Fluid systems based on ionic liquids for the pretreatment of lignocellulosic biomass
Abstract
Ionic liquids are substances with great potential for the development of better methods for the pretreatment of lignocellulosic biomass, which is a key step in the successful development of integral biorefinery schemes. Thus, different fluid systems based on ionic liquids were investigated for their utilisation in diverse process approaches: pairs of mutually immiscible ionic liquids, binary mixtures of a light alcohol with 1-ethyl-3-methylimidazolium acetate, and acetate ionic liquids based on a phosphonium cation. Fundamental knowledge was generated on these systems, and some of them were applied to the pretreatment of wood particles, achieving good degrees of fibrillation and decrystallisation under certain conditions.
Full text
TESE DE DOUTORAMENTO FLUID SYSTEMS BASED ON IONIC LIQUIDS FOR THE PRETREATMENT OF LIGNOCELLULOSIC BIOMASS María del Carmen Castro Valiña ESCOLA DE DOUTORAMENTO INTERNACIONAL PROGRAMA DE DOUTORAMENTO EN ENXEÑARÍA QUÍMICA E AMBIENTAL SANTIAGO DE COMPOSTELA 2018
DECLARACIÓNDAAUTORADATESE FLUIDSYSTEMSBASEDONIONICLIQUIDSFORTHEPRETREATMENTOF LIGNOCELLULOSICBIOMASS Dna. María del Carmen Castro Valiña Presento a miña tese, seguindo o procedemento adecuado ao Regulamento, e declaro que: 1) A tese abarca os resultados da elaboración do meu traballo. 2) No seu caso, na tese se fai referencia ás colaboracións que tivo este traballo. 3) A tese é a versión definitiva presentada para a súa defensa e coincide coa versión enviada en formato electrónico. 4) A tese non incorre en ningún tipo de plaxio de outros autores nin de traballos presentados por min para a obtención de outros títulos. SantiagodeCompostela,10deoutubrode2018 Asdo.: María del Carmen Castro Valiña
AUTORIZACIÓNDODIRECTOR/TITORDATESE FLUIDSYSTEMSBASEDONIONICLIQUIDSFORTHEPRETREATMENTOF LIGNOCELLULOSICBIOMASS Dna. Ana María Soto Campos e D. Héctor Rodríguez Martínez INFORMAN: QueapresentetesecorrespóndesecotraballorealizadoporDna.MaríadelCarmenCastro Valiña,baixoadireccióndeDna.AnaMaríaSotoCamposeD.HéctorRodríguezMartínez, eatitorizacióndesteúltimo,ea utorizamos asúa presentación ,considerando que reúneos r equisitos esixidosnoR egulamento deEstudosde DoutoramentodaUSC, e que comodirectoraedirector‐titordesta nonincorrenascausasde abstención establecidas naLei 40/2015. SantiagodeCompostela,10deoutubrode2018 Ana M. Soto Campos Héctor Rodríguez Martínez
To my parents Antonio and María del Carmen
ix Acknowledgements I would like to begin these lines by thanking Prof. Ana Soto, co-director of this doctoral thesis, for the opportunity to join the Research Group on Separation Processes and Phase Equilibria that she leads at the University of Santiago de Compostela (USC), and for all her guidance and share of knowledge. I am also indebted to Dr. Héctor Rodríguez, co-director and tutor of the thesis, for all his help and invaluable commitment, that have made this document possible. The work at his side, hand by hand, for almost four years has allowed me to develop as a professional and grow as a scientist; as well as to provide me with a taste for the small details, the rigor, and the work well done. My gratitude is extended to Dr. Eva Rodil, Dr. Óscar Rodríguez and, of course, Prof. Alberto Arce, always concerned about whether everything was going all right. I would also like to thank my laboratory colleagues, with whom I have shared great moments at different periods during these years. Borja, Iago, Marlen, and Raquel were the best mates I could have had in this scientific adventure. And a special mention to Iria, the lab mate who had to suffer me the most in the worst moments, thank you for always being there. Also, many thanks to Manuel and Oussama, who, during their short stays in the laboratory, made the long work days more enjoyable with their doses of optimism, happiness and friendliness. Part of the work in this thesis was carried out during a 3-month research stay at the University of Helsinki, in Finland. I am grateful to Prof. Ilkka Kilpeläinen for accepting me as a temporary member of his group during this time. It was a real pleasure for me to work under the supervision of Dr. Alistair King, who was always helpful and answered my questions very kindly. I have to make a mention to my colleagues in the laboratory, Ashley, Arno, Uula, and Niklas, with whom I spent great moments in the laboratory and enjoyed dinners and typical Finnish meals. And, of course, I cannot forget Daniel Rico and Queralt Martínez, who became my family in Helsinki during this period.
1. INTRODUCTION
1. Introduction 3 1. INTRODUCTION 1.1.Lignocelluloses:renewablefeedstockfora sustainableindustrialchemicalplatform 1.1.1.Context A current trend in the industry is the search for more sustainable processes, less polluting and with less environmental impact. At present, however, substances deriving from a non-renewable source still have a very high specific weight in most processes in the chemical industry, being uses as raw materials or as fuels. This is causing significant environmental problems, such as those related to global warming. To reverse these effects, a shift towards a new platform of industrial production of materials and chemicals based on renewable raw materials is necessary. The high volume of industrial production, together with the satisfaction of the standards demanded at present by the societies of the developed countries, and increasingly demanded by those of the developing countries, poses a significant challenge. An interesting approach to combine the high production volume and the required standards with the need of a sustainable chemical industry is the use of lignocellulosic biomass as raw material. Nature produces lignocelluloses in a biorenewable way in sufficient quantities to cover the demand of industrial production of the human being (Klemm et al., 1998). In addition, lignocelluloses also present the advantage of a more homogeneous geodistribution, in comparison to the fossil resources that today sustain most of the global productive scheme. And also importantly, it avoids the competition or interaction that other types of biomass would have with the food market (FitzPatrick et al., 2010). The plant cell walls in lignocellulosic biomass are complex structures mainly composed of three polymers: cellulose, hemicellulose, and lignin (Figure 1.1). The different chemical structures of the biopolymers give an idea of the chemical richness embedded in lignocelluloses, and their versatility to be the basis of an alternative chemical platform for the industrial production of a wide range of chemicals and materials. However, their appropriate exploitation in this regard is conditioned by their
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 4 recalcitrant character, which is the result of a natural evolution to resist degradation. With the three major polymers organised into complex non-uniform three-dimensional structures, this recalcitrance of lignocelluloses stems essentially from the crystallinity of cellulose, the hydrophobicity of lignin, and the encapsulation of cellulose by the lignin–hemicellulose matrix (Isikgor and Becer, 2015). Figure 1.1. Main components and structure of lignocelluloses. Reproduced from Isikgor and Becer (2015) by permission of the Royal Society of Chemistry. 1.1.2.Chemicalcompositionandstructuraldisposition Chemical composition The relative composition of cellulose, hemicellulose, and lignin in the cell wall of a plant varies according to its species, tissues, and maturity. Table 1.1 shows some examples
1. Introduction 5 of types of lignocellulosic biomass and their average content in the three biopolymers (Isikgor and Becer, 2015). In general terms, it can be claimed that lignocellulosic biomass consists of 35–50 % cellulose, 20–35 % hemicellulose, and 10–25 % lignin; with proteins, oils, and other organic and inorganic compounds making up the remaining fraction. Table 1.1. Some types of lignocellulosic biomass and examples of their typical content in cellulose, hemicellulose and lignin (adapted from Isikgor and Becer, 2015). Lignocellulosic biomass Cellulose (%) Hemicellulose (%) Lignin (%) Hardwood Poplar 51-53 26-29 15-16 Oak 40 36 24 Eucalyptus 54 18 22 Softwood Pine 42-50 24-27 20 Douglas fir 44 11 27 Spruce 456 23 28 Agricultural waste Wheat straw 35-39 23-30 12-16 Barley hull 34 36 14-19 Barley straw 36-43 24-33 6-10 Rice straw 29-35 23-26 17-19 Rice husks 29-36 12-29 15-20 Oat straw 31-35 20-26 10-15 Ray straw 36-47 19-25 10-24 Corn cobs 34-41 32-36 6-16 Corn stalks 35-40 17-35 7-18 Sugarcane bagasse 25-45 28-32 15-25 Sorghum straw 32-35 24-27 15-21 Grasses Gasses 25-40 25-50 10-30 Switchgrass 35-40 25-30 15-20 Cellulose is the most abundant biopolymer in Nature, with a basically structural function in vegetables. It is a carbohydrate formed by D-glucopyranose monomers linked together by β-1,4-O-glucosidic bonds (see chemical structure in Figure 1.1 – box labelled as ‘Cellobiose Unit’), resulting in the formation of long chains (Klemm et al., 2005; Wyman et al., 2005b). This chemical structure, together with its availability and high production rate (1011–1012 tonnes per year), confer a large potential to cellulose as a renewable raw material for a wide variety of products, including building block molecules for further chemical transformations, cellulosic polymers, and biofuels, among others (Klemm et al., 1998). Cellulose has a strong tendency to aggregate into highly ordered structures. This tendency is due to the formation of intraand intermolecular hydrogen bonds between the individual chains, which results in the packing of numerous cellulose chains in
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 6 crystalline structures known as fibrils (O’Sullivan, 1997). This crystalline structure is very compact since it is composed of a large number of hydrogen bonds at regular intervals (Klemm et al., 1998; Nagarajan et al., 2017). However, disordered amorphous regions together with highly ordered crystalline regions can be identified in the chains (Sjöström, 1993). Techniques such as X-ray diffraction and nuclear magnetic resonance spectroscopy have allowed the identification of four crystalline polymorphs of cellulose (O’Sullivan, 1997; Park et al., 2010): Cellulose I: It is the most abundant crystalline structure in Nature, and it is the one usually present in vegetables. There are two subtypes: cellulose I⍺ (produced by microbes, with a triclinic structure) and cellulose Iβ (typically present in higher plants, with a monoclinic structure) (Nagarajan et al., 2017). Cellulose II: It is usually obtained from cellulose I via treatments in basic media (mercerisation processes), or through solubilisation and subsequent recrystallisation; although in some specific cases native cellulose II could also be isolated (Kuga et al., 1993; O’Sullivan, 1997; Shibazaki et al, 1998). Cellulose III: This is a structure with two known varieties, cellulose IIII and cellulose IIIII, which are achieved after the ammonia treatment of cellulose I or cellulose II, respectively. It is a reversible transformation. Cellulose IV: It is achieved after subjecting cellulose III to a thermal treatment. Cellulose IVI or cellulose IVII can be obtained according to the starting variety of cellulose III. A scheme of the interconnection of crystalline structures described is shown in Figure 1.2. It must be noted that the crystal structures cellulose III and cellulose IV are of minor importance, since they are derived from cellulose I or cellulose II, and in addition they also have a lower stability. A comparison of the spatial configurations of the cellulose I and cellulose II polymorphs is presented in Figure 1.3. The native (biosynthesised) form cellulose I hosts a parallel chain strand arrangement, resulting in a compact structure of recalcitrant character; whereas cellulose II, thermodynamically more stable, presents its chains arranged in an anti-parallel fashion. The latter is a spatial configuration that confers a greater facility to react and, therefore, to be treated as raw material in
1. Introduction 7 industrial processes for its transformation and exploitation. Even greater reactivity is achievable with amorphous cellulose, in which the barriers presented by the crystalline structures are suppressed. Figure 1.2. Crystalline polymorphs of cellulose. Reproduced form Nagarajan et al. (2017) by permission of Elsevier. Hemicellulose is a flexible non-cellulosic polysaccharide composed of different sugars forming short and branched chains – see Figure 1.1. The sugars present in hemicellulose can be pentoses (xylose, arabionose) and hexoses (glucose, mannose, galactose), and they may be substituted to some extent with functionalities such as acetyl groups, hexuronic acids (glucuronic acid, methyl glucuronic acid, galacturonic acid) or deoxyhexoses (rhamnose, fucose). The main chain of hemicellulose can consist of a single unit (homopolymer), such as xylan; or of two or more units (heteropolymer), such as glucomannan. The branched nature of hemicellulose makes it amorphous and easier to hydrolyse than cellulose. Its main function is to provide the binding between cellulose and lignin (Sjöström, 1993). Lignin is a three-dimensional heteroamorphic polymer configured in a tridimensional network formed by monomers of phenylpropane (p-coumaryl,
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 8 p-coniferyl and p-sinapyl alcohols) – see chemical structures of these precursor units in Figure 1.1. The detailed structure of lignin is not exactly known since the methods used for its separation from the wood entail unavoidably its chemical or mechanical degradation (which gives an idea of the difficulty of isolating pure lignin from its native source within the lignocellulosic matrix). Hence, the "real" molecular weight of lignin is not known, although the literature suggests that lignins from hardwoods have a slightly lower molecular weight than those from conifers (Lin and Dence, 1992; Xie and Gathergood, 2013). Moreover, lignins isolated by means of different processes differ in chemical and physical properties. One of the most peculiar characteristics of lignin is that its monomers are linked by different types of covalent bonds that are irregularly distributed along the lignin chain, giving rise to a very complex structure, as can be seen in Figure 1.4. Figure 1.3. Comparison of intramolecular and intermolecular hydrogen bonds in cellulose I and cellulose II. Reproduced from Nagarajan et al. (2017) by permission of Elsevier. The main purpose of lignin is to provide the plant with rigidity, impermeability, and resistance to the attack by microorganisms and to oxidative stress (Hendriks and
1. Introduction 9 Zeeman, 2009). It is the most recalcitrant component of the plant cell wall, so that the greater the proportion of lignin, the greater the resistance to chemical and enzymatic degradations (Taherzadeh and Karimi, 2008). It can be considered as a “hydrophobic glue” that holds together the different lignocellulosic components. Figure 1.4. Generic molecular structure of softwood lignin highlighting various types of bonds. Reproduced from Zakzeski et al. (2010) by permission of the American Chemical Society. In addition to cellulose, hemicellulose, and lignin, lignocellulosic biomass contains other components in smaller proportions. Although these components are very different chemically, they can be classified into two categories: extractives and ashes. The extractives refer to a group of organic compounds of low molecular weight (including: terpenes, aliphatic and aromatic acids, alcohols, flavonoids, lignans, tannins, alkaloids, waxes, low molecular weight carbohydrates, etc.), which can be extracted from the biological material using water or organic solvents (Windeisen and Wegener, 2009; Rowell et al., 2013). Their main functions in the plant cell are those of external protection and as a reserve of nutrients. On the other hand, the ashes refer to the inorganic fraction of the material (mainly composed of inorganic salts of potassium, sodium, calcium, magnesium, and silica), which becomes ashes during its combustion.
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 16 biorefinery schemes must take into consideration the valorisation of all three biopolymers (FitzPatrick et al., 2010; Stöcker, 2008; Zakzeski et al., 2010). Figure 1.9 shows an overview of the potential products that could be obtained from cellulose, hemicellulose, and lignin in a lignocellulosic biorefinery according to this integral valorisation. Figure 1.8. Schematic comparison of two fundamental refinery concepts: petroleum refinery and biorefinery. Reproduced from Salan (2017) by permission of MedWin Publishers. As discussed in Section 1.1.3, the integral utilisation of biomass in this direction implies necessarily the disengagement of the biopolymers interconnected in the lignocellulosic matrix. Unfortunately, the separation of these biopolymers without degrading their chemical structures is highly challenging, likely being a reflection of the intrinsic resistance of the lignocellulosic matrix to its degradation. Inevitably this poses a major economic barrier for the development of a viable lignocellulosic biorefinery (Zhang et al., 2007), as the pretreatment methods developed to date or currently under development present a number of inconveniences and are still far from being satisfactory. An alternative approach, with the potential to lead to a more advantageous performance that the conventional methods, has started to emerge in the last years: the pretreatment of lignocelluloses with fluid systems based on ionic liquids.
1. Introduction 17 Figure 1.9. Potential scheme of products to be obtained from the different constituent biopolymers of lignocelluloses in a biorefinery. Reproduced from Kamm and Kamm (2004) by permission of Springer Nature. 1.2.Ionicliquids 1.2.1.Contextanddefinition Ionic liquids are salts that are liquid in their pure state at a low temperature. From a practical point of view, a mark of 373 K maximum for its melting (or glass transition) temperature is usually considered (Freemantle, 2010). For a salt to be liquid at such low temperatures, at least one of its constituent ions must be bulky or with a noticeable asymmetry, so that the charge is delocalised and the packing becomes more difficult, thus lowering the network energy of the crystalline form (Earle and Seddon, 2000). Although substances that adhere to the current definition of ionic liquid have been known for over a century, it was not until the last years of the 20th century that such substances and their potential were consolidated under a distinctive and attractive tag. These early years of the 21 st century have witnessed a sustained
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 18 increasing interest in ionic liquids, both in the academic environment and in industry. Figure 1.10 illustrates this growth through the evolution of the number of publications containing the concept ‘ionic liquid’ since the 1990s. Figure 1.10. Evolution, over the last 25 years, of the annual number of records retrieved by the Web of Science Core Collection when searching the topic “ionic liquid*”. (Search carried out in October 2018.) 1.2.2.Featuresandproperties The attractiveness of ionic liquids lies in their unique properties. Although it is difficult to generalise properties common to all ionic liquids due to their large number and variety (besides their ionic character and the limit in their melting or glass transition temperature, as imposed by the definition presented in Section 1.2.1), many of them often exhibit appealing sets of properties, including for example the following (Freemantle, 2010): Extremely low vapour pressure, essentially negligible under usual pressure and temperature operation conditions. Reasonably good thermal and chemical stabilities, leading frequently to a wide range of temperatures in which the ionic liquid is stable in the liquid state. Year 1993 1997 2001 2005 2009 2013 2017 Number of records 0 2000 4000 6000 8000 10000
1. Introduction 19 Low or negligible flammability. Great capacity to solvate a large number of varied compounds. In addition, the properties of ionic liquids can be 'customised' up to a certain level for a specific purpose, by judicious selection of the cation-anion combination and the design of the chemical structures of the constitutive ions (for example, by modifying the number or length of alkyl substituents) (Stark and Seddon, 2007). This set of properties stimulated the consideration of ionic liquids as alternative 'design' solvents in safer and environmentally friendlier processes (Seddon, 1997). However, the interest generated by ionic liquids and their singular properties goes beyond their use as solvents, and this is attested by the various industrial applications in which they have been gradually introduced (Freemantle, 2010, 2016; Maase, 2008; Plechkova and Seddon, 2008). An overview of their broad spectrum of applications, with a categorisation of their present level of development (from research and development to pilot plant scale, and to a commercialisation status), as evaluated by the German chemical company Iolitec GmbH, is provided in Figure 1.11. Figure 1.11. Selection of applications of ionic liquids and their development status, according to the company Iolitec GmbH (2017). STATUS:Research&development PilotplantCommercialised Functionalfluidsandadditives Lubricants Hydraulicfluids Additives Surfactants Processtechnology Gasseparation Biomassconversion Metalextraction Liquid‐liquidextraction Analytics SolventsforGCheadspace Proteincrystallisation GCmaterials Electrophoresis MatrixmaterialsforMALDI‐TOF‐MS Karl‐Fischertitration Synthesisandcatalysis Nanoparticlesynthesis Organicsynthesis Catalystsimmobilisation Enzymaticreactions Electrochemistry Sensors Supercapacitors Metaldeposition&electropolishing Dye‐sensitisedsolarcells Electrochromicwindows Fuelcells Batteries Heattransportandconversion Thermalfluids Sorptioncooling Phase‐changingfluids
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 20 1.3.Ionicliquidsystemsforthepretreatmentof lignocellulosicbiomass Swatloski et al. (2002) reported for the first time that some ionic liquids, such as 1butyl-3-methylimidazolium chloride, could truly dissolve cellulose without derivatising it, in high concentrations, and at relatively mild conditions. This result triggered a very active research on the topic, leading to the discovery of a good number of ionic liquids able to dissolve different types of cellulose at relevant levels (Badgujar and Bhanage, 2015; Mäki-Arvela et al., 2010; Muhammad et al., 2015; Pinkert et al., 2009; Wang et al., 2012). In this dissolution process, the interactions between the anion of the ionic liquid and the hydroxyl groups of cellulose play an important role: an anion with a good hydrogen bonding acceptor ability (such as chloride, acetate...) is a requirement for the ionic liquid to effectively dissolve cellulose. The role of the cation in this mechanism of dissolution is not that clear, although it has to be somehow actively involved. Deeper discussions on this can be found elsewhere (Badgujar and Bhanage, 2015; Wang et al., 2012). A few years after the ground-breaking discovery of ionic liquids with the capacity to dissolve cellulose, it was also found that lignocellulosic biomass (in particular, wood) could also be dissolved in ionic liquids (Fort et al., 2007, Kilpeläinen et al., 2007), under similarly mild conditions. In general terms, the solubilisation of lignocellulose seems to require similar properties of the ionic liquid than the solubilisation of cellulose (Brandt et al., 2013); so it is not surprising that, broadly speaking, ionic liquids that can dissolve cellulose are also capable of dissolving lignocelluloses (Badgujar and Bhanage, 2015; da Costa Lopes et al., 2013; Hou et al., 2017; Muhammad et al., 2015). Within this group of ionic liquids, there is a predominance of the chloride or acetate anions, mostly combined with dialkyl-substituted imidazolium cations. The solubility of lignin alone in ionic liquids was also investigated, mostly using various commercially available lignin models (e.g. kraft, alkaline, or organosolv) (Badgujar and Bhanage, 2015; Hou et al., 2017; Mäki-Arvela et al., 2010). It must be noted, however, that their solubilities may not be an adequate reflection of the corresponding behaviour of the lignin within a lignocellulosic material, since both structure and composition can be notably different (Hou et al., 2017). Regarding hemicellulose, it is more easily dissolved that either of the other major biopolymers in
1. Introduction 21 lignocelluloses, and to date no ionic liquids have been reported that can selectively dissolve hemicellulose but cannot dissolve cellulose and lignin (Hou et al., 2017). Given the negligible volatility of the ionic liquids and of the biopolymers/lignocelluloses, the regeneration of the latter from the ionic liquid solution has been most commonly practiced by addition of (molecular) solvents miscible with the ionic liquid and acting as antisolvents for the dissolved material, thus precipitating it out of the solution (da Costa Lopes et al., 2013; Rodríguez, 2016). The most widely antisolvent used is water, but unfortunately not much attention has been paid to the quantity of antisolvent really needed in an engineering context, nor to the significant energy penalty that may be involved in distilling off the antisolvent from its mixture with the ionic liquid for their recycling to the process (Rodríguez, 2016). For the case of dissolution of lignocellulosic biomass in ionic liquid, it has been shown that the design of an adequate scheme with the right antisolvents for selective precipitation at several stages can lead to a certain fractionation of the biopolymers (da Costa Lopes et al., 2013; Sun et al., 2009, 2011) – for example a cellulose-enriched material and isolated lignin (Figure 1.12). This is an evidence that ionic liquids can indeed break lignocellulosic bonds that hold the different constituent biopolymers together in the lignocellulosic matrix, without relying mainly in degradation of the lignin, as other pretreatment techniques do. Very interestingly, the crystallinity of cellulose in the regenerated lignocellulosic biomass fractions is lower than in the untreated material, with a change from cellulose I to cellulose II and a loss in fibrillar ordering that results in a higher amorphous component (Brandt et al., 2013); whereas the structures of lignin and hemicellulose remain essentially unaltered after treatment with ionic liquids (Wyman et al., 2009). All these observations would enable a more integral and efficient usage of the lignocellulosic feedstock in a biorefinery scheme (Sun et al., 2011). For example, for the production of biofuels such as bioethanol, enzymes can more efficiently hydrolyse into glucose the amorphous cellulose produced by means of the ionic liquid pretreatment than the highly crystalline cellulose as found in native lignocellulose (Dadi et al., 2007). Besides the approach based on the solubilisation of the entire lignocellulosic material in the ionic liquid, a second approach in the context of biomass pretreatment emphasises the chemical disruption of the chemical lignocellulose composite without achieving total dissolution in the ionic liquid (Brandt et al., 2013). For instance, the selective (partial) extraction of lignin from the lignocellulosic matrix can be carried out
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 22 with some ionic liquids able to dissolve lignin but not cellulose (Brandt et al., 2013; Hou et al., 2017). Figure 1.12. Flowchart for a process of dissolution of woody biomass in ionic liquid (IL) and its fractionated regeneration with an antisolvent strategy with water and acetone. Reproduced from Sun et al., 2011 by permission of the Royal Society of Chemistry. In either of the approaches, the utilisation of solvent systems combining ionic liquids with molecular solvents (e.g. water, acetone, aprotic polar solvents such as dimethylsulfoxide, etc.) has also been tested for a series of varied targets (Hou et al., 2017; Muhammad et al., 2015): the dissolution and processing of cellulose, the extraction of lignin from biomass, or the extraction of hemicellulose from paper-grade pulp, among many others. With this section providing only a brief and partial glance of what can be found in the very prolific literature on the subject, in light of the above, it can be stated that
1. Introduction 23 ionic liquids offer a basis for an extremely versatile technological platform for the pretreatment of lignocellulosic biomass. This versatility even includes the possibility of combining the ionic liquid technology with the conventional pretreatment methods already in place (Zhang et al., 2017). However, for ionic liquids to actually contribute to the fulfillment of the biorefinery paradigm through any of the multiple process variants investigated, there are still important challenges to be addressed (Brandt et al., 2013; Muhammad et al., 2015; Simmons et al., 2010; Zhang et al., 2017). A first group refers to the ionic liquids themselves: lowering their cost (as they are still comparatively too expensive), improvements in toxicity and biodegradability, or issues on thermal stability in the long term, are matters of concern. And of course, ionic liquids with better performance have to be sought as well, looking for e.g. a better tolerance to water in the system or a satisfactory pretreatment of biomass particles of bigger size (so that the large energy consumption in the preparation of the feedstock for the process can be reduced (Brandt et al., 2013)). Moving to challenges connected with the processing, the efficient recycling of the ionic liquid has been identified as a critical aspect in preliminary technoeconomic analyses (Binder and Raines, 2010; KleinMarcuschamer et al., 2011). This will require the effective removal of any remaining biomass fractions from the ionic liquid after the antisolvent addition (to avoid the buildup of unwanted pretreatment byproducts that could decrease performance (Simmons et al., 2010)); but also a minimised use of auxiliary substances such as antisolvents, along with a way of removing them from the ionic liquid with a tolerable energy input. All in all, ionic liquids with their unique properties constitute a very attractive family of substances for biomass processing in a biorefinery context. However, there is a need to overcome key challenges for the development and implementation of commercially viable processes, in which the entire process economy together with environmental and social impacts have to be properly optimised (Zhang et al., 2017). It is therefore a time for fascinating research on this topic!
2. OBJECTIVES
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 32 any proportion above 393 K (a temperature well below their limit of thermal stability (Cao and Mu, 2014; Keating et al., 2011)), but two liquid phases in equilibrium are generated for certain composition ranges below that temperature (Arce et al., 2007): the lower phase rich in the imidazolium ionic liquid and the upper phase rich in the phosphonium ionic liquid. In the context of a process for the pretreatment of lignocelluloses via dissolution with an ionic liquid, the use as antisolvent of a second ionic liquid showing switchable miscibility with the first one within a workable temperature range would have the benefit of maintaining integrally some of the characteristic advantages of ionic liquids (e.g. negligible vapour pressure) in the global solvent system. Moreover, only sensible heat would be transferred in the process of recovering the biomass-dissolving ionic liquid, as opposed to the much larger latent heat involved in the vaporisation of a classical antisolvent from its mixture with the ionic liquid. In this regard, this chapter will focus on the exploration of pairs of ionic liquids with the potential to exhibit spontaneous liquid-liquid demixing under certain conditions of temperature and composition, with only one of the members of the pair having a relevant capacity for dissolution of lignocelluloses. On the basis of the rather limited knowledge available to date on mutually immiscible ionic liquids and their liquid-liquid phase behaviour (Arce et al., 2006, 2007), their existence seems to be connected with the combination of ions with sufficiently dissimilar chemical structures. This structural difference may correspond to the cations, while the anion may be common to both ionic liquids. In fact, a common ion helps to reduce the complexity of the systems explored, facilitating the establishment of relations between the structural features of the involved ionic liquids and the produced phase behaviour. Thus, the combination of 1-alkyl-3-methylimidazolium chlorides or acetates (which are known to usually exhibit capacity for dissolution of lignocelluloses (Badgujar and Bhanage, 2015; Brandt et al., 2013)) with tetraalkylphosphonium or tetraalkylammonium ionic liquids with long alkyl substituents and the same anions is explored in this work, with rigorous determination of the liquid-liquid equilibrium in those cases showing any mutual immiscibility. A thermodynamic analysis complements the work, in order to get valuable information for a better understanding of the liquid-liquid equilibria in systems of mixed ionic liquids.
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 33 3.2.Theoreticalconsiderationsonliquid‐liquid equilibrium 3.2.1.Generalaspectsofliquid‐liquidequilibrium "Equilibrium" is defined as a situation in which no change occurs over time, although a real state of equilibrium is never reached, since it would take an infinite time. Therefore, it is considered that a system is in equilibrium when variations occur in very large intervals of time or are so small that their influence is negligible. From the Laws of Thermodynamics, it can be demonstrated that a closed system at constant temperature (T) and constant pressure (P) is in equilibrium when the total Gibbs free energy (G) of the system is minimal with respect to all possible changes at the given temperature and pressure (Smith et al., 2005). This condition can be mathematically expressed as: 0 , TP dG (3.1) where subscripts P and T indicated constant pressure and temperature respectively. This expression is valid regardless of the number of phases and components in the system, and it can be considered as a criterion or definition of equilibrium. For an open system with a single fluid phase, the Gibbs free energy is related to its independent variables, namely temperature, pressure, and number of moles (n) of each component i, through the following expression: iii dndTSdPVdG (3.2) where S and V are respectively the entropy and volume of the system, and μi is the chemical potential of component i. If a system with several components and several phases is considered, even if it is a closed system as a whole, each of the different phases is an open system that can exchange mass with the other phases, and Equation 3.2 is of application to each distinct phase: i k i k i kk dndTSdPVdG )()( )()( (3.3) where superscript k in parentheses refers to a generic phase. The sum of the expressions of Equation 3.3 applied to each of the phases will correspond to the total variation of Gibbs free energy in the system:
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 34 ki k i k idndTSdPVdG )()( (3.4) In order to consider the system in equilibrium, thermal and mechanical equilibria are needed, with T and P being uniform throughout all m phases in the entire system: )()2()1( m TTT (3.5) )()2()1( m P P P (3.6) Combining Equations 3.1, 3.4, 3.5, and 3.6, the following can be deduced: )()2()1( m iii (3.7) which constitutes the equilibrium criterion on the basis of the chemical potentials of each component in the mixture, for multiphasic systems with m phases. Unfortunately, the chemical potential is a rather abstract concept, not having an immediate equivalent in the physical world. Therefore, it is desirable to express the chemical potential in terms of magnitudes measurable in the laboratory. Thus, in order to facilitate a more practical application of the equilibrium criterion of Equation 3.7, the concept of fugacity is used. Defined by G. N. Lewis, fugacity is a function that satisfies the following equation (Prausnitz et al., 1999): 0 0ln i i ii f f TR (3.8) where R is the universal gas constant, and superscript 0 indicates a reference state. This reference state and its values of chemical potential (μi0) and fugacity (fi0) are arbitrary, although not independent, and as soon as one of them is set the other one is immediately fixed by Equation 3.8. By means of Equations 3.7 and 3.8, the equilibrium criterion can be expressed now as a function of fugacity, as follows: )()2()1( m iii fff (3.9) which is a more suitable expression from a practical point of view, and it is valid as long as all the reference states for all the phases are at the same temperature. Lewis also defined the activity of a component i (ai) as the ratio between its fugacity and the reference fugacity: 0 i i if f a(3.10)
3.Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 35 whi ch can be interpreted as how ‘active’ a substance is in relation to its reference potential at the state of interest and that at its reference state (Prausnitz et al., 1999). If the reference state for all phases is the same, the combination of Equations 3.9 and 3.10 yields the equilibrium criterion that depends on activities: )()2()1( m iii aaa (3.11) The activity of a species i is related to its composition by means of the activity coefficient i: i i ix a (3.12) where xi is the mole fraction of the species i. Hence Equation 3.11 can be transformed into: )()2()1( m iiiiii xxx (3.13) which is a popular and practical expression for the equilibrium criterion in multiphasic liquid systems. For a system of N components, each activity coefficient i(k) is a function of the temperature and pressure of the system and of the N-1 independent mole fractions in phase k. 3.2.2.Liquid‐liquidequilibriuminbinarysystems In an equilibrium system in which the only significant intensive variables are temperature, pressure, and chemical potential, and under the assumption that no chemical reaction takes place, the famous phase rule (credited to J. W. Gibbs) provides a means of determining the number of intensive variables that can be independently varied without changing the state of the system (Smith et al., 2005; Treybal, 1963): NF 2 (3.14) where F represents this number of independent variables (known as the number of degrees of freedom), is the number of phases, and N is the number of components. If a binary system (two components) in which two liquid phases coexist is considered, from application of Equation 3.14 it is clear that F = 2. In such binary system, each of the activity coefficients will be a function of temperature, pressure, and one independent mole fraction. Thus, if one of those variables is fixed, the liquid-liquid equilibrium of the binary system can be conveniently represented in the two-
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 36 dimensional plane by plotting the other two variables in the corresponding xand y-axes. For conditions of constant pressure (which is usually the case in the most relevant applications of liquid-liquid equilibrium at industrial level), the most typical representation is a temperature-composition diagram (also known as solubility diagram). Figure 3.1. Representative types of liquid-liquid equilibria for binary systems, represented in isobaric temperature-composition diagrams: a) “island” type; b) upper critical solution temperature (UCST) type; b) lower critical solution temperature (LCST) type. Reproduced from Smith et al. (2005) by permission of McGraw-Hill. Figure 3.1 shows three different types of liquid-liquid equilibria for binary systems, plotted in temperature-composition diagrams at constant pressure (Smith et al., 2005), with x1 in the x-axis being the mole fraction of species 1. In these diagrams, for a given temperature T, points A and B at the intersection of the corresponding horizontal line with the equilibrium curves (or binodal curves) determine the compositions in equilibrium x1 and x1 for the coexisting liquid phases (richer in species 2) and (richer in species 1). Figure 3.1a represents an “island”-type liquidliquid equilibrium, where the possibility of having two coexisting liquid phases in equilibrium is restricted to the temperature range between TU and TL, which are respectively the so-called upper critical solution temperature (UCST) and lower critical solution temperature (LCST). At temperature above TU or below TL, only one liquid phase will be obtained over the entire range of composition. “Island”-type liquid-liquid equilibrium behaviours occur rarely, as the binodal curves are usually interrupted by
3.Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 37 some other phase transition. If the binodal curves are interrupted by the freezing curve, only a UCST can exist (Figure 3.1b); whereas if they are interrupted by the bubble curve of the vapour-liquid equilibrium, only an LCST can exist (Figure 3.1c) (Smith et al., 2005). A large number of liquid pairs form systems without upper or lower critical solution temperatures, since a solid phase forms before the appearance of an LCST on cooling, and a vapour-liquid condition (vapour phase of the same composition and density as one of the liquid phases) occurs before the appearance of a UCST on heating (Treybal, 1963). 3.3.Experimental 3.3.1.Ionicliquids The ionic liquids 1-ethyl-3-methylimidazolium chloride ([C2mim]Cl) and 1-ethyl-3methylimidazolium acetate ([C2mim][OAc]) were purchased from Iolitec with nominal purities greater than 98 % and 95 %, respectively. 1-Butyl-3-methylimidazolium chloride ([C4mim]Cl), 1-methyl-3-octylimidazolium chloride ([C8mim]Cl), and 1-butyl3-methylimidazolium acetate ([C4mim][OAc]) were supplied by Fluka, with nominal purities greater than 95 %, 98 %, and 95 %, respectively. Trihexyl(tetradecyl)phosphonium chloride ([P6 6 6 14]Cl) was kindly donated by Cytec with a nominal purity greater than 97 %. 1-Hexyl-3-methylimidazolium chloride ([C6mim]Cl) was synthesised in-house by alkylation of 1-methylimidazole (Aldrich, >99 %) with 1-chlorohexane (Aldrich, 99,5 %), following an analogous procedure to that reported elsewhere for other 1-alkyl3-methylimidazolium chlorides (Bradley et al., 2002). Aliquat 336® (CAS number 63393-96-4) is the trade mark of a product corresponding to a mixture of trialkylmethylammonium chlorides, where the alkyl chains are octyl or decyl, with octyl predominating in a 2:1 mole ratio (Mikkola et al., 2006). Denominated here with the abbreviation [Aliquat]Cl, it was obtained from Sigma with a nominal purity in the range 85 %–95 % (with the impurities being mainly constituted by water and residual alcohols). The syntheses of trihexyl(tetradecyl)phosphonium acetate ([P6 6 6 14][OAc]) and Aliquat acetate ([Aliquat][OAc]) were conducted by a metathesis reaction between
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 38 potassium acetate (Sigma Aldrich, >99 %) and the corresponding chlorides of the desired cations, according to the procedure reported by Mikkola et al. (2015) for the synthesis of several tetraalkylphosphonium acetates. Both reactants were dissolved independently in ethanol (Panreac, 99.8 %), and then the alcoholic solutions were mixed and stirred overnight. The precipitate of potassium chloride was filtered off, and the ethanol was removed by rotary evaporation. After cooling to room temperature, acetone (Sigma Aldrich, 99.9 %) was added, the mixture was stirred and then placed in the freezer for 48 h. If precipitate was observed, the filtration stage, evaporation of acetone, and re-addition of fresh acetone was repeated; until no further precipitation was observed. All ionic liquids, regardless of being acquired from a commercial vendor or synthesised in-house, were purified by subjecting them to highly reduced pressure (<1 Pa) while being stirred and heated at ca. 330-340 K, to reduce the level of possible volatile impurities. Water is an impurity of particular concern, given the hygroscopic character of ionic liquids and the strong influence that it can have on their performance (Seddon et al., 2000). After the purification step, the final water content of the ionic liquids was measured by Karl-Fischer titration in a Metrohm 737 KF coulometer (Figure 3.2), and it was found to be 0.0023 in mass fraction for [P6 6 6 14][OAc], 0.0020 for [C2mim]Cl, 0.0012 for [C4mim][OAc], and in the range 0.0002-0.0008 for the rest of ionic liquids. Their chemical identity and the absence of significant levels of impurities in these purified products were confirmed by 1H and 13C nuclear magnetic resonance (NMR) spectroscopy analyses, run in a Varian Mercury 300 NMR spectrometer (Figure 3.2), or occasionally in a Bruker DRX-500 NMR spectrometer. The corresponding spectra are compiled in Appendix A. For the two ionic liquids prepared in-house by metathesis, the concentration of the spectator ions in the final products was measured, and found to be in the range 1600-1700 ppm for chloride (as determined by ion chromatography) and in the range 100-200 ppm for potassium (as determined by inductively coupled plasma optical emission spectroscopy, ICP-OES); which was considered acceptable for the purposes of this work. The chemical structures of the constitutive ions of the ionic liquids mentioned in this section are shown in Figure 3.3.
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 39 Figure 3.2. Equipment for the characterisation of the purified ionic liquids: Varian Mercury 300 NMR spectrometer (left) and Metrohm 737 KF coulometer (right). Figure 3.3. Chemical structures of the ions of the ionic liquids involved in this work: a) 1-alkyl-3methylimidazolium (R = ethyl, butyl, hexyl, or octyl), [C n mim]+ (where n stands for the number of carbon atoms in the alkyl substituent); b) methyltrioctylammonium, as the most representative chemical structure in [Aliquat]+ (see the main text for further details); c) trihexyl(tetradecyl)phosphonium, [P6 6 6 14]+; d) chloride, Cl-; e) acetate, [OAc]-; f) bis(trifluoromethylsulfonyl)amide, [NTf2]-. ab a) b) c) d) e) f) _
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 40 3.3.2.Thermalanalyses Thermal analyses to identify the range in which the ionic liquids remained thermally stable in the liquid state were carried out in order to determine the temperature ranges to be investigated in the liquid-liquid equilibrium studies. Thermal stability of the ionic liquids was determined by thermogravimetric analysis (TGA) using a TA Instruments Q500 thermogravimetric analyser with a weight precision of ±0.01 % (Figure 3.4). This apparatus was calibrated in weight with weights of certified mass, and in temperature with nickel of high purity (99.9945 %) by means of the determination of its Curie temperature, in accordance with the instructions by the manufacturer. An open platinum pan loaded with ca. 15-20 mg of sample was used in each case. The TGA runs were carried out at a heating rate of 5 K·min-1, from ambient temperature up to 773 K, using N2 (Praxair, 99.999 %) as balance purge gas and as sample purge gas (with flow rates of 40 mL·min-1 and 60 mL·min-1, respectively). Figure 3.4. TA Instruments Q500 thermogravimetric analyser. The lower end of the temperature range of the ionic liquids as stable liquids was determined by differential scanning calorimetry (DSC) using a TA Instruments Q2000 differential scanning calorimeter (Figure 3.5) with an RCS 90 refrigerated cooling system attached. The apparatus was calibrated with indium of high purity (99.99 %) by means of the determination of its onset melting temperature, in accordance with the
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 41 instructions by the manufacturer. Approximately 10-20 mg of each sample was placed in a 40 μL aluminium pan, sealed hermetically with a lid of the same material and loaded into the measuring chamber with an autosampler. An analogous, empty pan with its corresponding lid was used as reference. An initial heating at 5 K·min-1 up to 393 K was the initial step of the thermal program applied, followed by a 5-min isotherm at 393 K, and then two cooling-heating cycles in the temperature range 183-393 K, with cooling/heating ramps of 5 K·min-1 and the corresponding 5-min isotherms both at 183 K and at 393 K after finishing each cooling or heating ramp. A 50 mL·min-1 flow of N2 was used as sample purge gas. After ensuring that the DSC curves for the two full cooling-heating cycles were essentially coincident, the signal from the last cooling ramp was used for determination of (midpoint) glass transitions, and the signal from the last heating ramp was used for determination of (onset) melting temperatures. Figure 3.5. TA Instruments Q2000 differential scanning calorimeter. The analysis of all the TGA and DSC curves was made by means of the Universal Analysis 2000 software, version 4.5.0.5, by TA Instruments. 3.3.3.Identificationofmutuallyimmiscibleionicliquids Combinations of an imidazolium ionic liquid and a tetraalkylammonium or tetraalkylphosphonium ionic liquid, with either chloride or acetate as common anion,
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 48 determining the low end of the temperature range for experimental investigation of the liquid-liquid equilibrium in the system involving [C4mim]Cl. Table 3.2. Melting temperature (Tm) and/or glass transition temperature (Tg) for the ionic liquids involved in the liquid-liquid equilibria determined in this work, at atmospheric pressure. “Exp.” stands for experimental values determined herein, and “Lit.” stands for literature values. The values marked with an asterisk correspond to a Tg, while the other values correspond to a Tm. Ionic liquid Tm(K) and/or Tg(K) Exp. Lit.‡ [C2mim]Cl 357 362;a361b [C4mim]Cl 221* 343;b314;c222*, 321(m), 345(o)d,‖ [C6mim]Cl 208* 198*;c 223*d [C8mim]Cl 199* 186*;c 229*d [C2mim][OAc] 202* 195*;e 198*f [C4mim][OAc] 210* 203*;f203*g [Aliquat]Cl 323 253h [Aliquat][OAc] 250 251h [P6 6 6 14]Cl 207 203*;i216*j [P6 6 6 14][OAc] 206 not found ‡ References: a Ngo et al. (2000); b Kick et al. (2013); c Huddleston et al. (2001); d Diogo et al. (2013); e Troshenkova et al. (2010); f Guan et al. (2011); g Wei et al. (2015); h Mikkola et al. (2006); i CYTEC Industries Inc. (2011); j Pozo-Gonzalo et al. (2014). ‖ Polymorphs: (m), monoclinic; (o), orthorhombic. 3.4.2.Mutuallyimmisciblepairsofionicliquids The ionic liquid [Aliquat]Cl was combined with [Cnmim]Cl ionic liquids (n = 2, 4, 6, or 8). Liquid-liquid immiscibility was observed for the systems [C2mim]Cl + [Aliquat]Cl and [C4mim]Cl + [Aliquat]Cl, whereas the systems [C6mim]Cl + [Aliquat]Cl and [C8mim]Cl + [Aliquat]Cl were found to be totally miscible over the entire composition range and investigated temperature range. This is similar to what was previously reported for mixtures [Cnmim]Cl + [P6 6 6 14]Cl, which may give rise to liquid-liquid biphasic systems when the alkyl substituent chain of the imidazolium cation is pentyl or shorter, but they total miscibility at any composition for hexyl or longer substituents (Arce et al., 2006). Regarding the combinations of ionic liquids with acetate as common anion, [C2mim][OAc] and [C4mim][OAc] were combined with the acetates of the two tetraalkylpnictogenium cations, i.e. with [Aliquat][OAc] and with [P6 6 6 14][OAc]. Following the protocol described in Section 3.3.3, liquid-liquid biphases were found in the binary systems [C2mim][OAc] + [Aliquat][OAc] and [C2mim][OAc] + [P6 6 6 14][OAc],
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 49 but not in th e systems with [C 4mim][OAc]. Thus, in analogy with what has been described for the chloride-based systems in the paragraph above, an increase in the length of the alkyl substituent of the imidazolium ionic liquid leads to the disappearance of the liquid-liquid biphasic domain in the systems. A difference, however, is that in the case of the mixtures of acetates, a shorter alkyl substituent (butyl instead of hexyl) is sufficient for the total miscibility to occur. 3.4.3.Experimentalliquid‐liquidequilibriumdata For the above mentioned systems that showed distinct liquid-liquid biphasic character at some conditions of concentration and temperature, the liquid-liquid equilibrium was rigorously determined, according to the procedure described in Section 3.3.4. The mole fraction compositions of the phases in equilibrium for the binary systems [Cnmim]Cl + [Aliquat]Cl (n = 2 or 4), [C 2mim][OAc] + [Aliquat][OAc], and [C2mim][OAc] + [P6 6 6 14][OAc], at the different experimental temperatures tested, are reported in Table 3.3 (systems with chloride as a common ion) and Table 3.4 (systems with acetate as a common ion). Moreover, the liquid-liquid equilibrium data for the system [C2mim]Cl + [P6 6 6 14]Cl, which has already been reported in the literature (Arce et al., 2006) but only in graphical form, were also determined and are numerically reported in Table 3.3. From direct inspection of the tables, it can be observed that the lower phase is (very) rich in the imidazolium ionic liquid in all studied systems, whereas the upper phase is rich in the tetraalkylammonium/tetraalkylphosphonium ionic liquid. For a better analysis of the influence of different structural features of the ionic liquids on the liquid-liquid equilibria, the corresponding temperature-composition diagrams were built. Figure 3.8 shows a diagram of this type, displaying the liquidliquid equilibria of the systems [C2mim]Cl + [P6 6 6 14]Cl and [C2mim][OAc] + [P6 6 6 14][OAc]. For these two systems, it can be observed that the presence of phosphonium cations in the imidazolium-rich phase (the lower phase) is very small and even negligible, whereas the imidazolium cations are present in the phosphonium-rich phase (the upper phase) in a relevant concentration. The replacement of the chloride anion with the bigger acetate anion in these systems leads to greater concentrations of the imidazolium cation in the phosphonium-rich phase, increasing the degree of
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 50 miscibility (or, equivalently, decreasing the concentration range exhibiting immiscibility) at a given temperature. Table 3.3. Mole fractions of the imidazolium ionic liquid in the upper and lower phases (x1,up and x1,low, respectively) of the liquid-liquid equilibria of systems [Cnmim]Cl (1) + [Cation]Cl (2) (where n = 2 or 4, and [Cation]+ stands for [Aliquat]+ or [P 6 6 6 14]+), at the corresponding temperatures T and atmospheric pressure. T / K x 1,up x 1,low [C2mim]Cl (1) + [Aliquat]Cl (2) 358.2‡ 0.104 0.999 368.2 0.115 0.996 378.2 0.130 0.998 388.2 0.124 0.997 398.2 0.106 0.997 [C4mim]Cl (1) + [Aliquat]Cl (2) 358.2 0.377 0.969 368.2 0.468 0.949 378.2 0.491 0.952 388.2 0.493 0.954 398.2 0.483 0.951 408.2 0.480 0.951 [C2mim]Cl (1) + [P6 6 6 14]Cl (2) 373.2 0.098 1.000 383.2 0.093 1.000 393.2 0.087 1.000 403.2 0.084 1.000 413.2 0.086 1.000 423.2 0.080 1.000 ‡ This temperature is only slightly higher than the reported melting temperature experimentally determined for pure [C2mim]Cl (see Table 3.2), but it was included in the study in order to cover a sufficiently large temperature range, given the upper temperature limitation imposed by the thermal decomposition of [Aliquat]Cl (see Table 3.1). The influence of temperature in the system [C2mim]Cl + [P6 6 6 14]Cl is unclear, but in any case it is small over the studied temperature range, as the mutual solubility of these two ionic liquids can be taken as practically invariant in such range (within the experimental composition uncertainty). Contrarily, for the mixture of [C2mim][OAc] and [P6 6 6 14][OAc] a clear decrease in mutual solubility is observed as the temperature is risen. This suggests that the liquid-liquid equilibrium of the system [C2mim][OAc] + [P6 6 6 14][OAc] responds to an LCST-type phase behaviour. Interestingly, this is in contrast to the UCST-type behaviour previously observed in the literature (Arce et al.,
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 51 2006, 2007) for the system [C2mim][NTf2] + [P6 6 6 14][NTf2], with the same cations but with the much bulkier bis(trifluoromethylsulfonyl)amide ([NTf2]-) as common anion – see its chemical structure in Figure 3.3. The liquid-liquid equilibrium of the latter system is also represented in Figure 3.8 for direct visual comparison. Even though the mutual immiscibility for these pairs of ionic liquids is largely due to the great dissimilarity of their cations, Figure 3.8 evidences the importance that the nature of the common counterion (i.e. the anion) has in the type of liquid-liquid equilibrium generated. Table 3.4. Mole fractions of the imidazolium ionic liquid in the upper and lower phases (x1,up and x1,low, respectively) of the liquid-liquid equilibria of systems [C2mim][OAc] (1) + [Cation][OAc] (2) (where [Cation]+ stands for [Aliquat] + or [P 6 6 6 14]+), at the corresponding temperatures T and atmospheric pressure. T/ K x 1,up x 1,low [C2mim][OAc] (1) + [Aliquat][OAc] (2) 298.2 0.407 0.985 308.2 0.427 0.979 318.2 0.435 0.982 328.2 0.449 0.981 338.2 0.445 0.978 348.2 0.452 0.979 358.2 0.430 0.981 368.2 0.377 0.983 [C2mim][OAc] (1) + [P6 6 6 14][OAc] (2) 298.2 0.400 0.992 318.2 0.369 0.995 328.2 0.360 0.997 338.2 0.352 0.996 358.2 0.330 0.993 368.2 0.324 0.994 378.2 0.313 0.995 A different evolution of the liquid-liquid equilibrium with temperature is observed in Figure 3.9 for the systems [C2mim]Cl + [Aliquat]Cl and [C2mim][OAc] + [Aliquat][OAc]. In these cases, an initial increase in mutual miscibility occurs with an increase in temperature; but the trend switches at a certain temperature, and the immiscibility gap starts increasing with further increase of the temperature. The result is an hourglass-shaped system. This type of temperature-composition behaviour was previously reported for the liquid-liquid equilibrium of a binary system comprising an ionic liquid and a molecular solvent (Łachwa et al., 2006). However, the systems
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass ͷʹ ͵Ǥͻ Ǥ Figure 3.8. Temperature-composition diagram for the liquid-liquid equilibria of the systems [C2mim]Cl + [P6 6 6 14]Cl (solid inverted triangles) and [C2mim][OAc] + [P6 6 6 14][OAc] (solid triangles). Literature values for the system [C2mim][NTf2]+[P6 6 6 14][NTf2] (open black hexagons) are also represented (Arce et al., 2007). In the x-axis, the mole fraction of the corresponding imidazolium ionic liquid ( x imid ) is represented. The liquid-liquid biphasic domain corresponds to the region between the two branches of each system. Solid lines are shown as guides to the eye. Figure 3.9. Temperature-composition diagram for the liquid-liquid equilibria of the systems [C2mim]Cl + [Aliquat]Cl (squares) and [C2mim][OAc] + [Aliquat][OAc] (circles). In the x-axis, the mole fraction of the corresponding imidazolium ionic liquid ( x imid ) is represented. The liquid-liquid biphasic domain corresponds to the region between the two branches of each system. Solid lines are shown as guides to the eye. x imid T.
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter ͷ͵ ǡ ͵Ǥͻ͵ǤͺǤǡ ȋȏʹȐΪȏȐ ȏʹȐȏȐΪȏȐȏȐȌ ǡ ǦǤǡǡǦ Ǥ Figure 3.10. Temperature-composition diagram for the liquid-liquid equilibria of the systems [C2mim]Cl + [Aliquat]Cl (squares), [C2mim]Cl + [P6 6 6 14]Cl (inverted triangles), and [C4mim]Cl + [Aliquat]Cl (diamonds). In the x-axis, the mole fraction of the corresponding imidazolium ionic liquid ( x imid ) is represented. The liquid-liquid biphasic domain corresponds to the region between the two branches of each system. Solid lines are shown as guides to the eye. ͵ǤͳͲǦ ȏʹȐΪȏͳͶȐȏʹȐΪȏȐǡ ȀǤȏͳͶȐΪ ȏȐΪǡ ǢǤǡȏȐΪǡ ǡ ǮǦǯ ǦǤ ȏȐΪ ǦȏͳͶȐΪǤ x imid T.
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 54 account for the slight decrease observed in the immiscibility gap with [C2mim]Cl. A more marked effect can be observed in the evolution of the mutual miscibility with temperature, since there is practically no effect in the system with [P6 6 6 14]+, while the system with [Aliquat]+ exhibits the hourglass-shaped behaviour mentioned above. An analogous discussion could be applied to the comparison of the systems [C2mim][OAc] + [P6 6 6 14][OAc] and [C2mim][OAc] + [Aliquat][OAc] (not shown together in the same figure, but easy to compare by simultaneous consideration of Figures 3.8 and 3.9), although noting that in this case a variation from LCST behaviour to hourglass behaviour is actually observed when replacing the [P6 6 6 14]+ cation with the [Aliquat]+ cation. The liquid-liquid equilibrium for the system [C4mim]Cl + [Aliquat]Cl is also plotted in Figure 3.10, thus enabling a direct analysis of the influence of the length of the alkyl substituent chain of the imidazolium cation, via comparison with the system [C2mim]Cl + [Aliquat]Cl. In this case, the replacement of the ethyl substituent with a butyl substituent in the imidazolium cation causes a remarkable increase in the mutual miscibility, while maintaining the hourglass type behaviour. A further increase in the length of the substituent, to hexyl or octyl, will result in total miscibility of the two ionic liquids, over the entire temperature range investigated, as already described in Section 3.4.2. Therefore, the evolution of the liquid-liquid equilibrium in the systems [Cnmim]Cl + [Aliquat]Cl follows an analogous trend to what was previously reported for the systems [Cnmim]Cl + [P6 6 6 14]Cl (Arce et al., 2006), for which also gradually smaller immiscibility was observed with an increase in the alkyl substituent length, and from hexyl onwards the ionic liquids became completely miscible in any proportion. By simultaneous consideration of all the systems in Figure 3.10, it can be noted that the variation of the alkyl substituent length in the imidazolium ionic liquid has a much stronger influence in the liquid-liquid equilibria of the mutually immiscible ionic liquids than the variation of the lengths of the alkyl substituents (or the central atom) in the tetraalkylammonium/tetraalkylphosphonium cations. The reason for this may be the relative modification of the polarity caused by those variations. The butyl substituent in [C4mim]+ is sufficiently long (in contrast to the ethyl substituent in [C2mim]+) to create an ‘apolar tail’ in the cation, leading to a stronger chance of interaction with the ‘apolar shells’ of [P6 6 6 14]Cl or [Aliquat]Cl; hence increasing mutual miscibility. However, the replacement of [P6 6 6 14]+ with [Aliquat]+, although involving a change in the length of all four alkyl substituents, does not lead to so relevant
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 55 modifications in the charge distribution of the ion, as well as consequently in the interaction with the imidazolium ionic liquid. 3.4.4.Thermodynamicanalysis In the studied liquid-liquid equilibria, the ammonium/phosphonium ionic liquid barely enters the imidazolium-rich phase. Contrarily, the solubility of the imidazolium ionic liquid in the ammonium/phosphonium-rich phase is very significant. Thus, an interesting thermodynamic analysis of the systems can be made from the perspective of solution of the imidazolium ionic liquid in the ammonium/phosphonium ionic liquid. From the solubility data at several temperatures, the apparent enthalpy change of the solution process (Happ) can be obtained from the classical van’t Hoff equation: R H T xapp p up 1 ln ,1 (3.15) where x1,up stands for the mole fractions of the imidazolium ionic liquid in the ammonium/phosphonium-rich phase, measured at the absolute temperatures T; subscript p indicates constant pressure; and R is the universal gas constant. However, in order to reduce the propagation of errors and to better discriminate between true chemical effects and those effects due exclusively to statistical treatment in the analysis of the corresponding van’t Hoff plots, an approach proposed by Krug et al. (1976) has led to the common utilisation of the following modified version of Equation 3.15: R H TT xapp p hm up 11 ln ,1 (3.16) where Thm is the harmonic mean of the experimental temperatures, calculated as: n ii hm T n T 1 1(3.17) with n corresponding to the total number of experimental temperatures investigated. If a linear dependence is found when plotting ln(x1,up) against the difference (1/T – 1/Thm), Happ can be assumed as constant over the explored temperature range, and its value can be easily inferred from the slope of the linear fit. A non-linear behaviour implies that Happ changes with temperature in the studied interval, and in
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 56 this case a first approach to consider is to carry out a polynomial regression of order two (parabolic model) of the data (Mora et al., 2005): 2 ,1 1111 ln hmhm up TT c TT bax (3.18) where a, b and c are fit parameters. By derivation of this expression and comparison with Equation 3.16, the following equation is obtained for the calculation of Happ at each specific temperature: hm app TT cbRH 11 2(3.19) Figure 3.11 shows the modified van’t Hoff plots for the solution of the imidazolium ionic liquid in the ammonium/phosphonium ionic liquid for the liquidliquid equilibrium systems investigated in this work; as well as for the system [C2mim][NTf2] + [P6 6 6 14][NTf2], built from literature data (Arce et al., 2007), for comparative purposes. An analysis with the F-test (Devore, 2004) determined that the second-order term in Equation 3.18 was not statistically significant, at a significance level of 0.05, in the regression of the systems involving the phosphonium cation ([C2mim]Cl + [P6 6 6 14]Cl and [C2mim][OAc] + [P6 6 6 14][OAc]). Therefore, the data series of these two systems were fit to a straight line, and the numerical values of the slope and intercept are listed in Table 3.5., along with the root mean square deviation values (rmsd). The constant values of Happ for these two systems were derived from the corresponding slope values (via their multiplication by R), and are reported in Table 3.6. Table 3.5. Numerical values of the fit parameters of Equation 3.18 for the liquid-liquid equilibrium systems studied in this work (a zero value for parameter c is indicative of a linear fit). The corresponding root-mean square deviation (rmsd) values are also shown. System ab / K c / 106·K2 rmsd [C2mim]Cl + [Aliquat]Cl -2.07 -111 -9.75 0.027 [C4mim]Cl + [Aliquat]Cl -0.698 -506 -5.24 0.026 [C2mim]Cl + [P6 6 6 14]Cl -2.43 583 0 0.021 [C2mim][OAc] + [N8 8 8 1][OAc] -0.794 80.1 -1.31 0.026 [C2mim][OAc] + [P6 6 6 14][OAc] -1.05 335 0 0.005 [C2mim][NTf2] + [P6 6 6 14][NTf2] -0.820 -1139 0.769 0.027
3. Mutuallyimmiscibleionicliquidswithacommonanionofbasiccharacter 57 Figure 3.11. Modified van’t Hoff plots (natural logarithm of the mole fraction of imidazolium ionic liquid in the phosphonium-rich phase, x 1,up , versus the difference of inverses of the absolute temperature T and the harmonic mean temperature T hm ) for liquid-liquid systems comprising two mutually immiscible ionic liquids: [C2mim]Cl + [Aliquat]Cl (squares); [C4mim]Cl + [Aliquat]Cl (diamonds); [C2mim]Cl + [P6 6 6 14]Cl (inverted triangles); [C2mim][OAc] + [Aliquat][OAc] (circles); [C2mim][OAc] + [P6 6 6 14][OAc] (triangles); and [C2mim][NTf2] + [P6 6 6 14][NTf2] (open hexagons). Solid lines correspond to linear fits or to quadratic fits obtained with Equation 3.18. Table 3.6. Apparent enthalpy change (ΔHapp) and apparent entropy change (ΔSapp) for the solution of the imidazolium ionic liquid in the phosphonium ionic liquid in the systems [C2mim]Cl + [P6 6 6 14]Cl and [C2mim][OAc] + [P6 6 6 14][OAc]. System ΔHapp / kJ·mol-1 ΔSapp / J·mol-1·K-1 [C2mim]Cl + [P6 6 6 14]Cl -4.85 0.10 -32.4 0.4 [C2mim][OAc] + [P6 6 6 14][OAc] -2.79 0.01 -17.0 0.1 The apparent entropy change of solution (Sapp) could then be calculated by means of the rearranged Gibbs equation: T GH Sappapp app (3.20) where Gapp is the apparent Gibbs energy change of solution, which was evaluated from the experimental data according to the following expression: upapp xTRG ,1 ln (3.21) For systems with a linear behaviour, an option identified in the specialised literature is to evaluate Gapp at the specific temperature Thm, getting a single 103ꞏ(1/T - 1/Thm ) / K-1 -0.5 -0.3 0.0 0.3 0.5 ln x1,up -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 65 4. IONICLIQUID+ALCOHOLSYSTEMS. SOLUBILITYOFBIOPOLYMERS 4.1.Motivation Although ionic liquids often exhibit an appealing set of properties for their use as solvents in potentially sustainable processes (Freemantle, 2010), one of their most recurrent drawbacks is their relatively high viscosity, as compared to conventional molecular solvents. This has been found to impose important limitations in processes such as the dissolution of biopolymers in ionic liquids. For example, in the particular case of dissolution of cellulose in known cellulose-dissolving ionic liquids, higher levels of solubilisation have been experimentally achieved with increasing temperature, in spite of the fact that the process itself has been proven to be thermodynamically exothermic (and therefore it should be favoured at lower temperatures) (Andanson et al., 2015). The explanation for this contradiction has to reside in the kinetic limitation derived from the high viscosity of the ionic liquid medium. The use of molecular solvents as cosolvents of the ionic liquid in this kind of processes, assuming that they do not reduce significantly the dissolution capacity of the ionic liquid, might be of interest since the kinetics of the dissolution process would be facilitated as a result of the diminution of the viscosity (Stark and Seddon, 2007). This lowering in viscosity of the fluid medium would also be a concomitant benefit for the process from an engineering perspective. Moreover, the use of a cosolvent together with the ionic liquid would allow modulation of the solubility capacity by controlling the composition of the resulting solvent fluid, with the advantages that a fractionated solubility of the different biopolymers might have for some specific purposes. Of course, the cosolvent should be miscible with the biomass-dissolving ionic liquid and should also possess reasonably green credentials to fit within the general context of sustainability of the new process. On the other hand, since ionic liquids and lignocellulose biopolymers are nonvolatile, and if the lignocellulosic fractions or any derived solutes still in polymeric form are to be regenerated from the ionic liquid solution prior to any further utilisation, a logical approach for this regeneration is the use of a classical solvent as an antisolvent
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 66 to force precipitation. This antisolvent has to be miscible with the biomass-dissolving ionic liquid, while possessing a negligible ability to act as solvent of the solutes to be precipitated. Thus, upon addition of the antisolvent the solute will be precipitated out of the solution and will be recoverable from the medium by simple solid-liquid separation operations (e.g. by filtration); and the antisolvent will be removed from its mixture with the ionic liquid via vaporisation (e.g. by distillation). Due to the latter aspect, a characteristic of interest for the proposed antisolvent would be an intermediate volatility, for balancing two aspects: the energy required for its removal by vaporisation from the mixture with non-volatile ionic liquid for recycling of both substances to the process, and the safety and environmental risks associated with a too volatile compound. Additionally, the proposed antisolvent should again exhibit acceptably good green credentials that would not compromise the general sustainable character of the new process. In the regeneration of lignocelluloses from ionic liquid solution by addition of antisolvent, no emphasis has been put in the literature on the quantification of the antisolvent added to cause the precipitation of the solutes. This is a critical aspect in the conception and design of a process to be scaled up for real application at an industrial level. Also, water has been typically the antisolvent of choice in most of the literature available to date. Obviously, the green credentials of water perse are unbeatable; however, its high specific heat and relatively high boiling temperature pose an excessive energy penalty at the stage of recovering the ionic liquid from its mixture with the antisolvent by vaporisation of the latter. An appealing alternative to water in the above described role may be the use of light alcohols (methanol, ethanol, 1-propanol and 2-propanol), which present lower specific heats and boiling temperatures than water, thus enabling the possibility of a less energy intensive separation of the ionic liquid + antisolvent mixture. In addition, they can be considered to have reasonably good green credentials intrinsic to their nature (Henderson et al., 2011), and therefore they have the potential to lead to an improved environmental friendliness of the overall process, in spite of inconvenient characteristics such as their flammability. For the pretreatment of lignocellulosic biomass via dissolution, 1-ethyl-3methylimidazolium acetate ([C2mim][OAc]) has become an archetypical ionic liquid (Brandt et al., 2013), as it has shown a great capacity for the dissolution of different sources of lignocellulosic biomass and its main biopolymers (Sun et al., 2009, 2011). This ionic liquid has a series of favourable properties for industrial applications (Freire
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 67 et al., 2011): a relatively low viscosity (within the context of ionic liquids), low toxicity (with a value of LD50, the dose that would cause the death of 50 % of a group of test animals, greater than 2000 mg·kg-1), low corrosiveness, a liquid character far below room temperature, and an acceptably good thermal stability. Thus, this chapter concentrates on the study of the fluid systems constituted by [C2mim][OAc] and each of the four lightest alcohols, through their thermal characterisation and the rigorous determination of key thermophysical properties as a function of temperature and composition. These provide both a deeper knowledge for better understanding of the behaviour of the mixtures at a fundamental science level and critical information for an efficient design of an industrial process pretending to utilise these mixtures. The study of all four [C2mim][OAc] + alcohol systems allows for analysis of the effect of the length of the alkyl chain of the alcohol on the properties, as well as the position of the hydroxyl group in the case of the two propanols. The investigation of the solubility of representative standards of the major biopolymers of lignocellulosic biomass in mixtures of [C2mim][OAc] and alcohol is considered, in order to preliminary evaluate the potential of the alcohols as cosolvents of [C2mim][OAc] in processes involving a fractionated solubilisation of lignocellulose biopolymers. Complementary, to analyse the viability of using the alcohols as antisolvents of lignocellulose fractions previously dissolved in the ionic liquid, precipitation tests are also carried out. In fact, the possibility of having versatile substances capable of acting either as cosolvents at certain concentrations or as antisolvents at other concentrations would be highly attractive, as it would permit the fluid system to be operated in a continuous basis without having to totally vaporise the volatile compound from the ionic liquid for the recycling of the latter to the process. 4.2.Theoreticalconsiderationson thermophysicalproperties 4.2.1.Density,viscosity,refractiveindex,andsurface tension Density and viscosity are two fundamental properties in the characterisation of any fluid in the context of a chemical process. Surface tension is also a relevant property in the design of process units in which mass transfer between fluid phases plays a relevant
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 68 role. Complementarily, refractive index is a practical property which, being a fundamental physical property of substances, can be easily used to ascertain the composition of a mixture (particularly a binary mixture). Knowledge of these properties over the appropriate temperature and composition ranges for fluid mixtures based on ionic liquids is critically valuable for the design of real processes where they will be involved. The density ( ) of a substance is defined as its mass per unit volume, and is probably the most useful physical property. It does not only participate directly in many design calculations and simulations, but also in the calculation of many other properties (Riddick et al., 1986). One such property is the molar volume (V), which is defined for a pure compound as: M V(4.1) where M represents the molar mass of the compound. In the case of a mixture of different compounds, the use of a weighted average of their individual molar masses Mi (with the mole fractions xi as weighting coefficients) in the numerator of Equation 4.1 yields the molar volume of the mixture: ii Mx V(4.2) The dynamic viscosity ( ) of a fluid (often referred to as simply –and not unambiguously!– “viscosity”) relates to the friction resistance between its molecules, that limits its ability to flow. This resistance opposes the movement of particles on other adjacent particles, and it is considered as an internal friction of the molecules. It is caused by the attractive forces between the liquid molecules. Specifically, dynamic viscosity corresponds to the force per unit area necessary to maintain a unit velocity gradient between two parallel planes a unit distance apart (Riddick et al., 1986). It is defined via Newton’s law of viscosity: y vx yx (4.3) where yx is the force in the x direction on a unit area perpendicular to the y direction, vx is the component in the x direction of the velocity vector of the fluid, and acts as the proportionality constant. If Equation 4.3, with a constant value for , describes well the
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 69 resistance to flow of the fluid, independently of the magnitude of the shearing stress or of the velocity gradient, then such fluid is said to be Newtonian. The kinematic viscosity ( ) is defined as the ratio of the dynamic viscosity and the density of the fluid: (4.4) and it happens to be directly proportional to the time required for a liquid to flow down through a capillary tube under its own hydrostatic head (Riddick et al., 1986). Thus, it is the type of viscosity obtained directly in viscometers based on liquid efflux times through capillary tubes. If the density of the liquid is known, the dynamic viscosity can be easily calculated from the kinematic viscosity by means of Equation 4.4. The refractive index (n) of a substance is defined as the ratio between the velocity of light in vacuum and the velocity of light in the substance (Riddick et al., 1986). This dimensionless physical property is a function of the temperature of the medium and of the wavelength of the incident light. The effect of temperature on the refractive index of a liquid is mainly due to its influence on the degree of packing of the molecules of the liquid. Regarding the incident light, in the most common case the D1 and D2 lines of a sodium lamp are used, with a weighted mean wavelength of 589.26 nm (Riddick et al., 1986). In such chase, the determined refractive index is denoted as nD, and the dependency on the incident light does no longer apply. Density and refractive index are both present in the Lorentz-Lorenz expression for the calculation of the molar refraction (RM): M n n R D D M 2 1 2 2 (4.5) where all the variables have already been defined. The surface tension ( ) of a liquid can be defined as the force exerted in the plane of the surface per unit length (Poling et al., 2001). It is a measure of the cohesive forces between liquid molecules present at the surface (Tariq et al., 2012). 4.2.2.Excesspropertiesandpropertychangesofmixing Excess properties and property changes of mixing derived from thermophysical properties can provide valuable information on the behaviour of real mixtures.
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 70 The excess property of an extensive thermodynamic property (e.g. molar volume, internal energy, enthalpy, entropy, etc.) is defined as the difference between the real value of the property and the value of the property calculated for the same conditions of temperature, pressure, and composition by the ideal solution equations (Prausnitz et al., 1999). For a given property M, the mathematical expression is: idE M M M (4.6) where ME is the excess property, M is the value of the real property, and Mid is the property of the ideal mixture. For the particular case of molar volume (V), the value of Vid is calculated as the mole-fraction weighted sum of the molar volumes of the pure compounds (Vi): iii id VxV (4.7) The property change of mixing (ΔM), for a given property M, in a multicomponent mixture is defined as: iii MxMM (4.8) where M is the property of the solution, and xi and Mi are respectively the mole fraction and the property of the pure i-th component. Thus, for molar volume, viscosity, molar refraction, and surface tension the corresponding property changes of mixing, or deviation properties, can be written as: iii VxVV (4.9) iii x (4.10) iiMiMM RxRR , (4.11) iii x (4.12) For a property that may expand over several orders of magnitude for a given system, the direct calculation of the corresponding property change of mixing may provide little information of value. Such is the case of viscosity in systems involving an ionic liquid and a molecular solvent of low molar mass. For this property, instead, an analysis of the viscosity logarithm change of mixing (Δln(η/η0), with η0 being a reference viscosity equal to 1 in the units in which η is expressed):
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 71 iii x000 lnlnln (4.13) is preferred, especially if considering also that the viscosity of many fluid systems is quite often well described by a simple and classical mixing rule, that can be credited to Arrhenius and Kendall, which for a binary system would take the following form (Kendall and Monroe, 1917): 2211 lnlnln xx (4.14) By comparing Equation 4.9 with the combination of Equations 4.6 (applied to molar volume) and 4.7, it can be deduced that the expressions of VE and V are identical, and therefore their numerical values are always coincident. Developing the summations in the expressions of the excess molar volume for the case of a binary system, we get: 2211 VxVxVV E (4.15) where subscripts 1 and 2 refer to each of the components in the mixture. Similarly, from Equations 4.11-4.13, the following developed expressions for the viscosity logarithm change of mixing, the molar refraction change of mixing, and the surface tension change of mixing of binary systems can be obtained: 0 22 0 11 00 lnlnlnln xx (4.16) 2 2 1 1MMMM RxRxRR (4.17) 2211 xx (4.18) 4.2.3.Datacorrelation:influenceofthetemperature Properties of liquids such as density, viscosity, refractive index, or surface tension do generally tend to decrease with an increase with temperature. However, the pattern followed in each case for the evolution with temperature may be substantially different. For many liquids, the variation of density, refractive index and surface tension with temperature can be acceptably assumed to be linear over a relatively broad range of temperatures. Taa 10 (4.19) TbbnD 10 (4.20) Tcc 10 (4.21)
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 72 where T is the absolute temperature, and a0, a1, b0, b1, c0, and c1 are the fit parameters. In some other situations, however, the correlation by means of a polynomial expression of higher order may be more suitable. For the density of pure ionic liquids, for example, either a linear fit (Gu and Brennecke, 2002; Jacquemin et al., 2008; Deng et al., 2011) or a second order polynomial fit (Gomes de Azevedo et al., 2005; Jacquemin et al., 2007; Hasse et al., 2009) are usually adopted in the literature for the correlation with temperature: 2 210 TaTaa (4.22) where a2 is the fit parameter corresponding to the second-order term. The statistical Fisher’s F-test (Devore, 2004) can be a suitable tool to evaluate whether the addition of the quadratic term will be statistically significant or not in the polynomial correlation of these properties as a function of temperature. The evolution of the viscosity of liquids with temperature is markedly different. For many liquids, the temperature dependency of the dynamic viscosity over wide temperature ranges can be suitably correlated by means of the Arrhenius-type equation proposed by Andrade (1930), for example in the case of simple solvents such as light alcohols. Its expression is: TR Ea exp (4.23) where R is the universal gas constant, T is the absolute temperature, and η∞ (“viscosity at infinitive temperature”) and Ea (“activation energy”) are the fit parameters. However, the 2-parameter Andrade equation is often unable to provide a good description of the evolution with temperature of the viscosity of other liquids, such as glass-forming liquids, and typically ionic liquids. For the latter, the use of the VogelFulcher-Tammann (VFT) equation (Vogel, 1921; Fulcher, 1925; Tammann, 1926), with three fit parameters, is preferred. In its modified version by Cohen and Turnbull (1959), the VFT equation is expressed as: 0 5.0 exp TT k TA (4.24) where A, k, andT0 are the fit parameters. In this expression, thanks to different theoretical rationales (Cohen and Turnbull, 1959; Adam and Gibbs, 1965), a physical meaning can be attributed to T0: it can be considered as an ideal glass transition temperature, i.e. a temperature below which the fluid exists as an equilibrium glass
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 73 where the mass-transporting motions are frozen out (Angell and Moynihan, 1969). Although it cannot be achieved in finite time scale experiments due to kinetic reasons, it should be slightly lower than the experimentally obtained glass transition temperature Tg (Gibbs and DiMarzio, 1958). Since the dynamic viscosity values for a fluid can span several orders of magnitude in the experimental temperature range explored, a fairer comparison of the evaluation of the quality of the fits provided by Equations 4.23 and 4.24 is better carried out on the basis of the relative standard deviation (SDrel) instead of the regular standard deviation, according to the following expression: 5.0 2 1, ,exp, 1 dat n icalci calcii pdat rel nn SD (4.25) where subscripts “exp” and “calc” refer to the experimental and calculated values respectively, ndat is the total number of experimental data points in the correlated series, and np is the number of fit parameters in the correlating equation. 4.2.4.Datacorrelation:influenceofthecomposition The description of a physical property of a liquid mixture as a function of composition can be conducted through the adequate correlation of the corresponding excess property or property change of mixing. One of the most common equations used for this purpose is the empirical Redlich-Kister polynomial (Redlich and Kister, 1948), which adopts the following mathematical form for binary mixtures: m k k kxxAxxQ 0 2121 (4.26) where Q is the excess property or property change of mixing, x1 and x2 are the mole fractions of the components of the mixture, and Ak are the m + 1 polynomial coefficients to be fit (with m representing the degree of the resulting polynomial). The quality of the fit obtained in each case can be evaluated by means of the root mean square deviation (rmsd), which is defined as: 5.0 1 2 ,exp, 1 rmsd n icalcii zz n(4.27)
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 80 cryogenic thermostat, with an uncertainty of 0.1 K. Series of twelve consecutive measurements were carried out for each sample, with the first two measurements being systematically disregarded (to allow adequate stabilisation of the liquid surface after contact of the plate and the sample, as well as homogenisation of the wetting of the plate by the sample). The values reported were the average of the remaining ten measurements. An uncertainty of 0.3 mN·m-1 was typically observed. Figure 4.4. Atago RX-5000 refractometer (left) and Anton Paar Abbemat 500 refractometer (right). Figure 4.5. Krüss K11 tensiometer (left) and detail of the platinum ‘plate’ folded in a cylindrical shape to perform measurements with smaller amounts of sample (right).
4.Ionicliquid+alcoholsystems.Solubilityofbiopolymers 81 4.3.5.Solubilitymeasurementsandprecipitationtests The solubilities of MCC, xylan, and Indulin AT in binary mixtures of [C2mim][OAc] and alcohol (as well as in the corresponding pure solvents) were determined by gradual addition and dissolution of controlled amounts of the polymer (weighing each added amount in the Mettler Toledo AE240 analytical balance previously mentioned in Section 4.3.2) to the liquid solvent. A mixture of known composition of [C2mim][OAc] + (methanol or ethanol) was initially placed in a jacketed glass cell (Figure 4.6), connected to an Ultraterm-200 P Selecta thermostatic water bath to maintain the temperature constant to within ±0.1 K throughout the experience. A first addition of biopolymer standard was carried out, with the content of the cell being vigorously stirred for 4-12 h. The stirring in these experiments was performed either by magnetic stirring with a Teflon-coated stirring bar or, in those cases where too high viscosities were achieved, by mechanical stirring with a metal rod coupled to an IKA RW 16 Basic overhead stirrer (Figure 4.6). After ceasing the stirring, the complete dissolution of the added solute was inspected by direct visual observation. Stepwise additions of biopolymer were subsequently repeated in each case, followed by the corresponding stirring period and assessment of dissolution, until the solubility limit was reached. This was typically manifested by undissolved solute at the bottom of the cell, suspended particles, or turbidity in the liquid phase. In the case of solubility of Indulin AT, occasionally very dark solutions were obtained, which prevented the confirmation of solubilisation by simple visual observation. In such situations, the quantification of the dissolved lignin content was carried out by measurement of the absorbance of the mixtures conveniently diluted with a 0.1 N aqueous solution of NaOH, at a wavelength of 360 nm, in an Agilent Technologies 8453 UV-visible spectrophotometer (Figure 4.7), following a similar procedure to that reported by Lee et al. (2009). The solubility limit would correspond to the lowest concentration for which a plateau of constant absorbance is observed, although in those cases with a particularly high ratio of ionic liquid to alcohol in the solvent, the high solubilisation of the lignin rendered the solution difficult to handle and no absorbance measurement could be carried out suitably. A microscopic analysis was performed, by means of a Leica DMRE7 optical microscope (Figure 4.8), in order to check the nature of the solubilisation of some of the samples investigated. Moreover, for a specific experiment on selective dissolution of a
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 82 mixture of the three aforementioned polymers by mixtures of [C2mim][OAc] + methanol, FT-IR spectra were carried out in a Varian FT-IR 670 spectrometer (Figure 4.9), using KBr pellets and recording a total of 32 scans in the wavelength range 500-4000 cm-1. Figure 4.6. Jacketed glass cell used for solubility experiments, with stirring via a metallic rod attached to an IKA RW 16 Basic overhead stirrer. Figure 4.7. Agilent 8453 Technologies UV-visible spectrophotometer.
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 83 Figure 4.8. Leica DMRE7 optical microscope. Figure 4.9. Varian FT-IR 670 spectrometer. On the basis of the solubility results of binary system [C2mim][OAc] + methanol, precipitation tests to evaluate the ability of methanol as antisolvent were carried out at room temperature. Mixtures of 3 g of [C2mim][OAc] and 0.15 g of the biopolymer standard (or a combination of polymer standards) were placed in capped glass vials and stirred magnetically until total dissolution. Controlled amounts of methanol were
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 84 gradually added. After each addition the resulting system was stirred for 30-45 min and allowed to settle for several minutes. 4.4.Resultsanddiscussion 4.4.1.Thermalcharacterisationof[C2mim][OAc]+ alcoholsystems Either in its role as cosolvent or antisolvent, it is desirable to have a possibility of easily removing the light alcohol from its mixtures with [C2mim][OAc], to recover the pure ionic liquid as needed. In principle, the vaporisation of the alcohol from the non-volatile ionic liquid at moderate temperatures could be a preferred strategy. To confirm that this is possible for the [C 2mim][OAc] + (methanol, or ethanol, or 1-propanol, or 2propanol) systems, TGA experiments were carried out for samples covering the composition range of the binary systems studied. Figure 4.10 shows the TGA thermograms for the pure [C2mim][OAc] and for its mixtures with methanol, ethanol, 1-propanol, or 2-propanol over the entire composition range. No TGA runs were carried out for the pure alcohols due to their totally volatile character. For pure [C2mim][OAc] a one-step decomposition curve was obtained, with an onset decomposition temperature (Td) of 475 K (see Figure B.15 in Appendix B). This value is in reasonably good agreement with the values of 489, 492, and 494 K respectively reported in the literature by Clough et al. (2013), Zhao et al. (2012), and Cao and Mu (2014), taking into account that our TGA runs were carried out at half the heating rate than theirs (5 K·min-1 versus 10 K·min-1). Almeida et al. (2012a) also reported 470 K as the temperature at which significant weight loss for this ionic liquid started to occur in their TGA experiments. The more conservative value Td,5%, which provides a better guidance for the maximum temperature at which the ionic liquid can be operated in practice (Clough et al., 2013; Smiglak et al., 2006), was calculated from the TGA curve obtained herein, and it was found to be 427 K (see Figure B.15 in Appendix B).
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 85 Figure 4.10. TGA thermograms for the binary systems [C2mim][OAc] + alcohol, from pure [C2mim][OAc] (top, right) to a 0.10 mole fraction of [C2mim][OAc], at a step composition of 0.10 in mole fraction. Solid and dashed lines are alternatively used for facilitation of the identification of each thermogram. Alcohol: a) methanol, b) ethanol, c) 1-propanol, d) 2-propanol. Regarding the mixtures of [C2mim][OAc] with alcohols (methanol, ethanol, 1propanol, or 2-propanol), all thermograms present a similar pattern (Figure 4.10). Starting at room temperature and with increasing temperature, there is a weight loss from the beginning as a result of the inherent volatility of the alcohol. Next, there is a horizontal inflection point in the curve (transition from convex to concave), occurring at a temperature Tip, and the decomposition proceeds thereafter in a similar way to that observed for the pure ionic liquid. Table 4.2 shows that there is an acceptable correspondence between the sample weight percent remaining at the aforementioned horizontal inflection point (%wtip) and the mass fraction of the ionic liquid (wIL) in the sample (perhaps with the exception of the samples with the highest concentration of alcohol, for which the correspondence is worse, likely due to higher losses by evaporation during the taring of the TGA balance prior to the start of the heating ramps of the runs). This suggests that the weight loss in the low-temperature part of the T / K 300 350 400 450 500 550 600 Weight / % 0 25 50 75 100 T / K 300 350 400 450 500 550 600 Weight / % 0 25 50 75 100 T / K 300 350 400 450 500 550 600 Weight / % 0 25 50 75 100 T / K 300 350 400 450 500 550 600 Weight / % 0 25 50 75 100 ab cd
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 86 thermogram (below Tip) is due to vaporisation of the alcohol, whereas the weight loss in the high-temperature part (above Tip) is due to the decomposition of the ionic liquid. Table 4.2. Mole fraction (xIL) and mass fraction (wIL) composition of ionic liquid in the binary mixtures of [C2mim][OAc] + (methanol, ethanol, 1-propanol, or 2-propanol) for the TGA experiments, along with the temperature (Tip) and remaining sample weight percent (%wtip) for the low-temperature (from convex to concave) inflexion points of the thermograms, and the corresponding pseudo 5 % onset decomposition temperatures (T’d,5%). x IL w ILTip / K %w t ip T ’ d,5%/ K [C2mim][OAc] + methanol 0.1000 0.3712 419 46.2 448 0.2000 0.5705 406 65.6 438 0.3000 0.6948 403 73.8 438 0.4000 0.7798 413 79.4 445 0.5000 0.8416 413 84.2 445 0.6000 0.8885 415 88.1 444 0.7000 0.9253 419 92.0 444 0.8003 0.9551 418 94.7 451 0.9003 0.9796 415 97.2 448 [C2mim][OAc] + ethanol 0.1000 0.2910 387 44.8 430 0.2000 0.4802 410 52.2 442 0.3000 0.6130 418 64.0 445 0.4000 0.7112 410 72.8 442 0.4999 0.7869 412 79.2 445 0.5998 0.8470 413 84.7 446 0.7001 0.8961 409 89.4 446 0.7990 0.9362 413 93.1 447 0.9000 0.9708 420 96.1 449 [C2mim][OAc] + 1-propanol 0.1017 0.2427 401 26.3 436 0.2004 0.4152 398 45.2 436 0.3001 0.5485 400 57.6 436 0.3998 0.6535 411 65.3 445 0.4997 0.7388 409 73.5 443 0.5996 0.8092 412 79.7 444 0.6998 0.8685 408 83.7 441 0.8002 0.9190 407 88.9 443 0.8991 0.9619 410 92.1 444 [C2mim][OAc] + 2-propanol 0.1000 0.2394 403 30.0 440 0.2000 0.4145 417 46.1 448 0.3000 0.5483 421 59.7 450 0.4000 0.6537 425 66.7 455 0.5000 0.7390 428 74.2 458 0.5998 0.8093 423 80.9 456 0.7000 0.8686 422 86.8 455 0.7993 0.9186 426 91.9 460 0.9000 0.9622 426 95.3 458
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 87 With the exception of the system with 2-propanol, the Tip values are clearly lower than the Td,5% of [C2mim][OAc] (427 K). This is indicative of the possibility of achieving a practically total vaporisation of the alcohol (methanol, ethanol, or 1-propanol) in its mixtures with the ionic liquid by a simple heating ramp up to Tip, without causing concomitant thermal degradation of the ionic liquid. Conversely, for the system [C2mim][OAc] + 2-propanol it is observed that the Tip values of most of the tested samples (in particular above a certain threshold concentration of ionic liquid in the mixture) are rather similar to the Td,5% calculated for pure [C2mim][OAc]; so the recovery of pure ionic liquid from the mixture via the heating ramp procedure could be more difficult. Since the specific heats and enthalpies of vaporisation of the alcohols (2.54 J·g-1·K-1 and 1.17 kJ·g-1 for methanol, 2.44 J·g-1·K-1 and 0.92 kJ·g-1 for ethanol, and 2.39 J·g-1·K-1 and 0.79 kJ·g-1 for 1-propanol; all values at 298 K (Riddick et al., 1986)) are largely lower than those of water (4.18 J·g-1·K-1 and 2.44 kJ·g-1, at 298 K (Riddick et al., 1986)), the use of these alcohols as antisolvents for the precipitation of lignocellulosic biopolymers from ionic liquid solution can be expected to notably reduce, in comparison to the use of water, the energy cost associated with recovering the ionic liquid by vaporisation of the molecular antisolvent. Nevertheless, in order to carry out this recovery by e.g. distillation, the accurate knowledge of vapour-liquid equilibria of the mixtures [C2mim][OAc] + alcohol should be determined for validation of the alleged savings of energy and viability of the recovery process. In this regard, the vapour-liquid equilibrium data reported by Cai et al. (2011) and Li et al. (2012) on ternary systems [C2mim][OAc] + alcohol + alkyl acetate look promising, since they point to little effect on the boiling temperature of the alcohols at low concentrations of the ionic liquid in the system. However, at high concentrations of ionic liquid in the system, the distillation of the alcohol could be harder, and appropriate experimental work in addressing this issue should be considered. Above Tip, the fraction of sample remaining may be assumed to be constituted essentially by the ionic liquid, with some traces of the alcohol. If the horizontal tangent to the TGA curve at Tip is taken as the baseline reference for the calculation of an onset decomposition at the high-temperature part of the curve, a pseudo 5 % onset decomposition temperature (T’d,5%) for the ionic liquid can be determined. The values of T’d,5% thus calculated are shown in Table 4.2. They lie in the approximate range 430460 K for all binary samples, being somewhat higher than the Td,5% found for the pure
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 88 ionic liquid. Although the difference is too small as to unambiguously state that there is a stabilising effect of the alcohol on the ionic liquid, it seems clear that the presence of the alcohol does not have a detrimental effect on the thermal stability of [C2mim][OAc] as compared to that of the ionic liquid in neat. Since the TGA runs carried out in this work are dynamic experiments, the temperature thresholds of thermal stability must be taken with care. Lower temperature limits might be considered if dealing with process plants operating in the midor long-term. However, the values reported in this section can be taken as fair estimations of the actual limiting temperatures at which industrial units in real processes could be operated. A comparison of the TGA curves of the four binary systems studied, at a given composition, namely for an ionic liquid mole fraction of 0.70, is provided in Figure 4.11. In this figure, a trend can be observed in the thermograms of the systems with primary alcohols (methanol, ethanol, 1-propanol), for which the mass loss upon heating increases with an increase in the alkyl chain length of the alcohol. This seems to correlate well with the simple fact that, as the molar mass of the alcohol is increased, the ionic liquid mole fraction of 0.70 is equivalent to a gradually lower ionic liquid mass fraction: 0.93 in the mixture with methanol; 0.90 in the mixture with ethanol; and 0.87 in the mixture with either of the propanols. However, a somewhat dissimilar behaviour is observed for the system with 2-propanol, which exhibits a relatively enhanced thermal stability. This suggests the presence of alternative/additional ways of interaction of the molecules of 2-propanol with the ions of the ionic liquid. The determination of the stable liquid range of the investigated systems was completed with a complementary study of thermal events in the mixtures by DSC. Figures 4.12 shows that no thermal events were observed for the different mixtures above ca. 200 K, thus confirming all four systems remain liquid at least down to such low temperature (approximately the lowest reliable temperature of the apparatus used), throughout the entire composition ranges. Therefore, no risk of crystallisation should be expected when using the studied mixtures in a process at temperatures even a far way below conventional ambient temperature.
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 89 Figure 4.11. TGA thermograms for the binary systems [C2mim][OAc] + alcohol, with a composition of 0.70 in mole fraction. Alcohol: methanol (magenta, long-dashed line), ethanol (brown, dot-dash line), 1-propanol (green, short-dashed line), or 2-propanol (blue, dotted line). The thermogram for pure [C2mim][OAc] (black, solid line) is also depicted for visual reference. Figure 4.12. Stacked DSC thermograms (heating ramps) for the binary system [C2mim][OAc] + alcohol, from pure [C2mim][OAc] (top) to pure alcohol (bottom) at a step composition of 0.10 in mole fraction. Alcohol: a) methanol, b) ethanol, c) 1-propanol, d) 2-propanol. Solid and dashed lines have been alternatively used to facilitate identification of each thermogram. T / K 300 350 400 450 500 550 600 Weight / % 0 25 50 75 100 T / K 200 220 240 260 280 300 320 Heat flow per unit mass of sample a T / K 200 220 240 260 280 300 320 Heat flow per unit mass of sample T / K 200 220 240 260 280 300 320 Heat flow per unit mass of sample T / K 200 220 240 260 280 300 320 Heat flow per unit mass of sample b dc
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 96 Figure 4.15. Density ρ for the binary system [C2mim][OAc] + 1-propanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.5.) Solid lines correspond to the correlation by means of a second-degree polynomial. Figure 4.16. Density ρ for the binary system [C2mim][OAc] + 2-propanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.6.) Solid lines correspond to the correlation by means of a second-degree polynomial. T / K 275 295 315 335 355 / gꞏcm -3 0.7 0.8 0.9 1.0 1.1 1.2 T / K 275 295 315 335 355 / gꞏcm -3 0.7 0.8 0.9 1.0 1.1 1.2
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 97 Table 4.7. Fit parameters (a0, a1, a2) of the quadratic fit of the density of the binary systems [C2mim][OAc] + (methanol, ethanol, 1-propanol, or 2-propanol) for its correlation as a function of temperature, at different mole fractions of ionic liquid (xIL), by means of Equation 4.22. The corresponding standard deviations (SD) are also shown. xILFit parameters and standard deviation ‡ a0 a1 104 a2 107 SD [C2mim][OAc] + methanol 0.0000 1.03615 -7.27936 -3.64286 0.00002 0.1000 1.14356 -8.57142 1.01429 0.00001 0.2000 1.27591 -13.8225 10.6071 0.00088 0.3000 1.22945 -8.52334 2.57143 0.00000 0.4000 1.25211 -8.43048 2.79286 0.00000 0.5000 1.26682 -8.23413 2.75000 0.00000 0.6000 1.27680 -7.99414 2.57143 0.00000 0.7000 1.27873 -7.40994 1.80000 0.00001 0.8003 1.28527 -7.27827 1.74286 0.00000 0.9003 1.31203 -8.57096 4.00714 0.00006 1.0000 1.32415 -8.93278 4.63571 0.00004 [C2mim][OAc] + ethanol 0.0000 0.97231 -3.92673 -7.88452 0.00006 0.1000 1.07893 -6.66002 -1.64405 0.00003 0.2000 1.14305 -7.67354 0.81071 0.00001 0.3000 1.18602 -8.02204 1.86905 0.00001 0.4000 1.21652 -8.12168 2.36429 0.00000 0.4999 1.23922 -8.05301 2.50714 0.00000 0.5998 1.25518 -7.82805 2.34762 0.00001 0.7001 1.26104 -7.19943 1.56905 0.00002 0.7990 1.27456 -7.16539 1.60357 0.00002 0.9000 1.29686 -7.82047 2.75000 0.00005 1.0000 1.31034 -7.99955 3.06190 0.00004 [C2mim][OAc] + 1-propanol 0.0000 0.94212 -1.56821 -10.7714 0.00009 0.1017 1.04992 -5.24234 -3.55119 0.00004 0.2004 1.11726 -6.94495 -0.20119 0.00002 0.3001 1.16250 -7.60661 1.26429 0.00001 0.3998 1.19596 -7.84051 1.92857 0.00001 0.4997 1.22111 -7.81397 2.11548 <0.00001 0.5996 1.24306 -7.80302 2.28214 <0.00001 0.6998 1.26085 -7.71062 2.28690 <0.00001 0.8002 1.27492 -7.53289 2.13690 0.00001 0.8991 1.27815 -6.83106 1.16667 0.00003 1.0000 1.30025 -7.36225 2.05833 0.00001 [C2mim][OAc] + 2-propanol 0.0000 0.86245 2.97206 -19.1429 0.00007 0.1000 1.00119 -2.97437 -7.56607 0.00004 0.2000 1.12865 -8.48619 2.09643 0.00013 0.3000 1.14395 -7.20131 0.64286 0.00001 0.4000 1.18261 -7.63239 1.64821 <0.00001 0.5000 1.21264 -7.78044 2.11607 <0.00001 0.5998 1.23567 -7.71625 2.19464 <0.00001 0.7000 1.25469 -7.58652 2.13750 <0.00001 0.7993 1.26630 -7.17374 1.61250 0.00001 0.9000 1.27905 -6.94639 1.36071 0.00002 1.0000 1.29981 -7.33391 2.01250 0.00001 ‡ Units: a0 and the standard deviation, in g·cm-3; a1, in g·cm-3·K-1; a2, in g·cm-3·K-2.
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 98 The quadratic fit adopted here is in contrast to the linear fit adopted by QuijadaMaldonado et al. (2012) for the density of pure [C2mim][OAc] in their study of the binary system [C2mim][OAc] + ethanol. A direct comparison of their density data and those reported herein for this specific system are possible at the temperatures of 298.15 K and 328.15 K. Figure 4.17 depicts such comparison, in which a very good agreement is observed between the two datasets at each indicated temperature. Figure 4.17. Comparison of the experimental density data ( ρ ) reported for the binary system [C2mim][OAc] + ethanol in this work (solid symbols) and those reported by Quijada-Maldonado et al. (2012) (open symbols), as a function of the mole fraction of ionic liquid ( x IL ), at atmospheric pressure and at the temperatures 298.15 K (circles) and 328.15 K (squares). A visual evolution of the viscosity with the temperature, at different constant compositions over the entire composition range for the binary systems studied, is provided in Figures 4.18 to 4.21. As it can be easily observed, the evolution of this property with the temperature, obeying a typical exponential decay, is starkly different than that described above for the density. From an application perspective in a chemical process context, it is obvious that low temperatures should be avoided especially with streams rich in the ionic liquid, due to their high viscosity. For the correlation of the experimental data of dynamic viscosity with temperature, two equations were tested: the classical Arrhenius-like equation proposed by Andrade (1930), which is known to provide a good description of the viscosity of simple solvents such as alcohols (Equation 4.23); and the Vogel-Fulcher-Tammann (VFT) equation (Equation 4.24), more x IL 0.0 0.2 0.4 0.6 0.8 1.0 / gꞏcm -3 0.7 0.8 0.9 1.0 1.1 1.2
4.Ionicliquid+alcoholsystems.Solubilityofbiopolymers 99 appropriate for the viscosity of substances that tend to form glasses, such as ionic liquids. The values obtained for the fit parameters of Equations 4.23 and 4.24 after correlation of all data series at a fixed composition for all the binary systems studied, along with the relative standard deviations (calculated by means of Equation 4.25), are reported in Table 4.8. For all compositions and systems, by comparison of the SDrel values, the VFT equation led to a clearly better description of the experimental data for the pure ionic liquid and most of its mixtures with any of the alcohols. However, attempts to correlate the series corresponding to the pure alcohols and the mixtures with xIL=0.10 with the VFT equation led to abnormally low values of the T0 parameter (with the exception of the system involving methanol). For this reason, only the Andrade equation was used in these cases. Thus, the fits shown as solid lines in Figures 4.18 to 4.21 correspond to the VFT equation for all series with xIL=0.20 or higher, while they correspond to the Andrade equation for the series with xIL=0.10 or xIL=0.00 (pure alcohol). An exception to this was the series with xIL=0.10 for the system [C2mim][OAc] + methanol, which was also carried out with the VFT equation. A good agreement between the experimental viscosity data and these correlations can be qualitatively observed throughout. Quijada-Maldonado et al. (2012) also investigated the viscosity of mixtures of [C2mim][OAc] and ethanol. The comparison of their data and experimental data presented herein at 298.15 K and at 328.15 K is provided in Figure 4.22. In spite of a general good agreement, the literature data are somewhat lower than those reported herein, especially at 298.15 K and at high concentrations of the ionic liquid. This may be due to a lower water content of the ionic liquid used in the present work. Figures 4.23 to 4.26 show the experimental refractive index for the binary systems studied, as a function of temperature, for different compositions. The linearity of the variation of the refractive index with temperature was evaluated by means of an F-test analysis, which yielded a statistically negligible quadratic term for practically all series at a significance level of 0.05 (as opposed to what was found for density). Therefore, a lineal polynomial expression (Equation 4.20) was used to fit the data of each composition series of all binary systems studied. The corresponding fit parameters and standard deviations are listed in Table 4.9, and the resulting straight lines are plotted along with the experimental data in Figures 4.23 to 4.26.
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 100 Figure 4.18. Viscosity for the binary system [C2mim][OAc] + methanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.3.) The inset plot provides a detailed view of the low viscosity range (below 100 mPa·s). Solid lines correspond to the correlation by means of the VFT equation ( x IL 0.10) or the Andrade equation ( x IL = 0.00). Figure 4.19. Viscosity for the binary system [C2mim][OAc] + ethanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.4.) The inset plot provides a detailed view of the low viscosity range (below 100 mPa·s). Solid lines correspond to the correlation by means of the VFT equation ( x IL 0.20) or the Andrade equation ( x IL 0.10). T / K 275 285 295 305 315 325 / mPaꞏs 0 100 200 300 400 500 600 700 T / K 275 285 295 305 315 325 / mPaꞏs 0 20 40 60 80 100 T / K 275 295 315 335 355 / mPaꞏs 0 100 200 300 400 500 600 700 T / K 275 295 315 335 355 / mPaꞏs 0 20 40 60 80 100
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 101 Figure 4.20. Viscosity for the binary system [C2mim][OAc] + 1-propanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.5.) The inset plot provides a detailed view of the low viscosity range (below 40 mPa·s). Solid lines correspond to the correlation by means of the VFT equation ( x IL 0.20) or the Andrade equation ( x IL 0.10). Figure 4.21. Viscosity for the binary system [C2mim][OAc] + 2-propanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.6.) The inset plot provides a detailed view of the low viscosity range (below 40 mPa·s). Solid lines correspond to the correlation by means of the VFT equation ( x IL 0.20) or the Andrade equation ( x IL 0.10). T / K 275 295 315 335 355 / mPaꞏs 0 50 100 150 200 250 300 T / K 275 295 315 335 355 / mPaꞏs 0 10 20 30 40 T / K 275 295 315 335 355 / mPaꞏs 0 50 100 150 200 250 300 T / K 275 295 315 335 355 / mPaꞏs 0 10 20 30 40
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 102 Table 4.8. Fit parameters (η∞, Ea, A, k, T0) of the Andrade equation (Equation 4.23) and of the VFT equation (Equation 4.24) for the correlation of viscosity as a function of temperature in the binary systems [C2mim][OAc] + (methanol, ethanol, 1-propanol, or 2-propanol), at different mole fractions of [C2mim][OAc] (xIL). The corresponding relative standard deviations (SDrel), calculated according to Equation 4.25, are also shown. xIL Andrade equation ‡ VFT equation ‡,‖ η∞x 103 EaSDrel A x 103 kT0SDrel [C2mim][OAc] + methanol 0.0000 9.074 -10.09 0.00 - - - - 0.1000 10.09 -12.25 0.01 1.223 975.0 66.07 0.008 0.2000 11.93 -13.64 0.01 2.383 911.6 84.06 0.001 0.3000 5.835 -16.97 0.02 4.936 718.2 125.2 0.002 0.4000 21.06 -20.94 1.16 5.141 758.9 136.4 0.000 0.5000 0.546 -25.63 0.04 6.068 740.8 151.9 0.001 0.6000 0.119 -30.64 0.05 5.365 796.4 158.2 0.001 0.6999 0.012 -37.69 0.09 5.942 775.2 171.1 0.002 0.8003 0.002 -43.02 0.13 9.829 674.3 185.9 0.006 0.9003 0.000 -49.76 0.15 8.905 702.2 190.4 0.004 1.0000 0.000 -55.54 0.26 8.590 708.6 194.6 0.001 [C2mim][OAc] + ethanol 0.0000 4.142 -13.77 0.01 - - - - 0.1000 8.161 -13.95 0.00 - - - - 0.2000 8.585 -15.34 0.01 1.316 1205 66.76 0.000 0.3000 3.802 -18.77 0.04 3.055 894.2 116.5 0.016 0.4000 2.036 -21.62 0.04 3.736 894.2 127.6 0.001 0.4999 0.531 -26.10 0.07 5.161 808.8 148.8 0.002 0.5998 0.066 -32.52 0.17 6.077 747.0 167.4 0.029 0.7001 0.019 -36.68 0.19 7.039 747.0 173.7 0.006 0.7990 0.002 -42.72 0.31 8.136 719.6 183.2 0.005 0.9000 0.000 -49.33 0.46 8.460 715.6 189.8 0.005 1.0000 0.000 -54.93 0.63 8.459 711.4 194.4 0.002 [C2mim][OAc] + 1-propanol 0.0000 1.380 -17.98 0.00 - - - - 0.1017 3.611 -17.26 0.00 - - - - 0.2004 5.008 -17.65 0.01 0.812 1418 65.29 0.007 0.3001 3.387 -19.82 0.02 1.295 1267 90.48 0.005 0.3998 1.812 -22.53 0.04 3.118 959.0 129.4 0.002 0.4997 0.717 -25.93 0.06 4.182 877.8 147.5 0.002 0.5996 0.256 -29.49 0.08 4.503 867.0 157.2 0.001 0.6998 0.074 -33.53 0.12 5.609 809.7 170.0 0.003 0.8002 0.021 -37.49 0.17 6.017 795.7 177.4 0.005 0.8991 0.005 -41.76 0.23 5.002 837.5 180.1 0.009 1.0000 0.001 -46.42 0.33 7.945 725.0 193.4 0.002 [C2mim][OAc] + 2-propanol 0.0000 0.268 -22.13 0.00 - - - - 0.1000 1.268 -20.08 0.00 - - - - 0.2000 2.512 -19.55 0.01 0.798 1329 83.12 0.002 0.3000 2.017 -21.32 0.02 1.566 1145 107.2 0.001 0.4000 1.059 -24.09 0.03 2.899 962.7 133.5 0.001 0.5000 0.411 -27.53 0.05 4.168 866.4 151.8 0.002 0.5998 0.159 -30.91 0.06 2.895 997.4 149.1 0.003 0.7000 0.037 -35.44 0.10 6.952 744.4 177.8 0.003 0.7993 0.012 -39.04 0.13 6.983 755.6 182.2 0.003 0.9000 0.003 -43.12 0.18 8.251 719.0 190.0 0.004 1.0000 0.001 -46.79 0.23 8.029 722.8 193.6 0.001 ‡ Units: ∞, in mPa·s; Ea, in kJ·mol-1; A, in mPa·s·K-0.5; k and T0, in K. ‖ Abnormally low values of T0 were obtained when attempting to correlate with the VFT equation the data series with xIL = 0.00 (pure alcohol) and xIL = 0.10. These correlations were discarded.
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 103 Figure 4.22. Comparison of the experimental dynamic viscosity data ( η ) reported for the binary system [C2mim][OAc] + ethanol in this work (solid symbols) and those reported by QuijadaMaldonado et al. (2012) (open symbols), as a function of the mole fraction of ionic liquid ( x IL ), at atmospheric pressure and at the temperatures 298.15 K (circles) and 328.15 K (squares). Figure 4.23. Refractive index ( n D ) for the binary system [C2mim][OAc] + methanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.3.) Solid lines correspond to linear fits. x IL 0.0 0.2 0.4 0.6 0.8 1.0 / mPaꞏs 0 20 40 60 80 100 120 140 160 T / K 275 285 295 305 315 325 n D 1.28 1.32 1.36 1.40 1.44 1.48 1.52
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 104 Figure 4.24. Refractive index ( n D ) for the binary system [C2mim][OAc] + ethanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.4.) Solid lines correspond to linear fits. Figure 4.25. Refractive index ( n D ) for the binary system [C2mim][OAc] + 1-propanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.5.) Solid lines correspond to linear fits. T / K 275 295 315 335 355 n D 1.32 1.36 1.40 1.44 1.48 1.52 T / K 275 295 315 335 355 n D 1.32 1.36 1.40 1.44 1.48 1.52
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 105 Figure 4.26. Refractive index ( n D ) for the binary system [C2mim][OAc] + 2-propanol, at atmospheric pressure and as a function of temperature ( T ), at different mole fractions of [C2mim][OAc] ( x IL ): , 0.00; , 0.10; ▲, 0.20; , 0.30; , 0.40; □, 0.50; ■, 0.60; , 0.70; ▼, 0.80; ○, 0.90; ●, 1.00. (For a greater degree of precision of the mole fractions, please refer to Table 4.6.) Solid lines correspond to linear fits. The evolution of the surface tension of mixtures of [C2mim][OAc]+(methanol, ethanol, 1-propanol, or 2-propanol) with temperature at different compositions covering the entire composition range is shown in Figures 4.27 to 4.30. Analogously to what was described for the refractive index, the apparent linear trend observed for the surface tension with temperature for all binary systems studied was confirmed by means of the F-test, and a straight line was used to correlate the experimental data of surface tension with temperature (Equation 4.21). Least-squares regressions led to the fit parameters listed in Table 4.9. The quality of the proposed fits for correlation of the experimental data can be graphically corroborated in Figures 4.27 to 4.30. T / K 275 295 315 335 355 n D 1.32 1.36 1.40 1.44 1.48 1.52
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 112 Table 4.12. Excess molar volume (VE ), viscosity logarithm change of mixing (ln( / 0)), molar refraction change of mixing (RM ), and surface tension change of mixing ( ) for the binary system [C2mim][OAc] + 1-propanol, at atmospheric pressure and different temperatures (T), as a function of the mole fraction of [C2mim][OAc] (xIL). xILT/ K 288.15 298.15 308.15 318.15 328.15 338.15 348.15 V E/ cm3·mol-1 0.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.1017 -0.511 -0.565 -0.628 -0.704 -0.794 -0.903 -1.033 0.2004 -0.618 -0.686 -0.765 -0.860 -0.975 -1.114 -1.283 0.3001 -0.651 -0.722 -0.804 -0.903 -1.022 -1.166 -1.340 0.3998 -0.647 -0.714 -0.792 -0.887 -1.000 -1.136 -1.299 0.4997 -0.606 -0.669 -0.739 -0.824 -0.925 -1.046 -1.190 0.5996 -0.539 -0.593 -0.652 -0.723 -0.809 -0.911 -1.032 0.6998 -0.437 -0.481 -0.526 -0.582 -0.649 -0.729 -0.823 0.8002 -0.313 -0.347 -0.377 -0.417 -0.463 -0.518 -0.582 0.8991 -0.144 -0.180 -0.199 -0.219 -0.244 -0.272 -0.306 1.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 ln(η/η0) 0.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.1017 0.183 0.238 0.279 0.316 0.347 0.378 0.405 0.2004 0.211 0.295 0.365 0.430 0.489 0.524 0.563 0.3001 0.254 0.350 0.426 0.491 0.547 0.582 0.625 0.3998 0.293 0.386 0.461 0.522 0.573 0.618 0.656 0.4997 0.312 0.396 0.461 0.513 0.555 0.593 0.625 0.5996 0.297 0.370 0.428 0.468 0.499 0.528 0.552 0.6998 0.273 0.328 0.370 0.401 0.425 0.440 0.457 0.8002 0.202 0.243 0.273 0.299 0.305 0.316 0.325 0.8991 0.111 0.139 0.157 0.147 0.155 0.164 0.168 1.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 RM 0.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.1017 0.016 0.026 0.015 0.024 0.029 0.023 0.076 0.2004 0.027 0.027 0.025 0.036 0.034 0.030 0.072 0.3001 0.045 0.034 0.037 0.048 0.050 0.054 0.065 0.3998 0.053 0.035 0.046 0.045 0.069 0.049 0.059 0.4997 0.052 0.029 0.035 0.036 0.035 0.037 0.051 0.5996 0.047 0.031 0.032 0.032 0.047 0.035 0.049 0.6998 0.056 0.029 0.031 0.037 0.029 0.040 0.041 0.8002 0.048 0.019 0.011 0.023 0.020 0.023 0.031 0.8991 -0.007 -0.001 -0.013 -0.044 -0.015 -0.073 -0.086 1.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 σ / mN·m-1 0.0000 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.1017 -0.5 -0.6 -0.1 -0.6 -0.5 0.0 -0.1 0.2004 -1.2 -1.6 -1.4 -1.6 -1.0 -0.8 -0.7 0.3001 -1.3 -1.3 -1.0 -0.8 -0.5 -0.1 0.3 0.3998 -1.7 -1.8 -1.3 -1.5 -0.9 -0.5 -0.2 0.4997 -2.1 -2.1 -1.8 -1.9 -1.7 -1.2 -1.2 0.5996 -2.6 -2.9 -2.7 -3.2 -2.8 -2.5 -2.8 0.6998 -2.8 -2.8 -2.7 -2.8 -2.6 -2.2 -2.8 0.8002 -2.7 -2.6 -2.6 -2.5 -2.7 -2.4 -2.3 0.8991 -2.0 -2.0 -2.0 -2.0 -2.0 -1.7 -1.5 1.0000 0.0 0.0 0.0 0.0 0.0 0.0 0.0
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 113 Table 4.13. Excess molar volume (VE), viscosity logarithm change of mixing (ln( / 0)), molar refraction change of mixing (RM), and surface tension change of mixing ( ) for the binary system [C2mim][OAc] + 2-propanol, at atmospheric pressure and different temperatures (T), as a function of the mole fraction of [C2mim][OAc] (xIL). xIL T/ K 288.15 298.15 308.15 318.15 328.15 338.15 V E/ cm-3·mol-1 0.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.1000 -0.544 -0.626 -0.727 -0.851 -1.001 -1.184 0.2000 -0.641 -0.703 -0.829 -0.987 -1.182 -1.421 0.3000 -0.622 -0.727 -0.856 -1.018 -1.216 -1.459 0.4000 -0.632 -0.732 -0.853 -1.004 -1.189 -1.415 0.5000 -0.616 -0.705 -0.813 -0.946 -1.110 -1.308 0.5998 -0.566 -0.643 -0.734 -0.846 -0.983 -1.148 0.7000 -0.482 -0.545 -0.616 -0.704 -0.810 -0.939 0.7993 -0.344 -0.396 -0.445 -0.507 -0.581 -0.670 0.9000 -0.192 -0.228 -0.255 -0.289 -0.328 -0.375 1.0000 0.000 0.000 0.000 0.000 0.000 0.000 ln(η/η0) 0.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.1000 0.238 0.303 0.360 0.409 0.448 0.488 0.2000 0.240 0.345 0.439 0.520 0.586 0.650 0.3000 0.295 0.411 0.511 0.592 0.664 0.728 0.4000 0.346 0.456 0.548 0.625 0.688 0.747 0.5000 0.376 0.472 0.545 0.612 0.664 0.715 0.5998 0.369 0.458 0.519 0.568 0.602 0.638 0.7000 0.354 0.403 0.441 0.479 0.504 0.529 0.7993 0.278 0.313 0.340 0.364 0.378 0.391 0.9000 0.174 0.188 0.195 0.212 0.216 0.223 1.0000 0.000 0.000 0.000 0.000 0.000 0.000 RM 0.0000 0.000 0.000 0.000 0.000 0.000 0.000 0.1000 0.026 0.020 0.027 0.045 0.079 0.059 0.2000 0.307 0.313 0.319 0.328 0.346 0.357 0.3000 0.075 0.061 0.065 0.081 0.088 0.099 0.4000 0.081 0.086 0.074 0.093 0.127 0.118 0.5001 0.082 0.079 0.216 0.077 0.078 0.109 0.5998 0.102 0.077 0.075 0.094 0.147 -0.019 0.7000 0.102 0.078 0.071 0.078 0.083 0.112 0.7994 0.076 0.054 0.059 0.069 0.063 0.068 0.9000 0.062 0.039 0.039 0.045 0.039 0.058 1.0000 0.000 0.000 0.000 0.000 0.000 0.000 σ/ mN·m-1 0.0000 0.0 0.0 0.0 0.0 0.0 0.0 0.1000 -0.6 -0.6 -0.6 -0.5 -0.7 -0.1 0.2000 -1.3 -1.3 -1.1 -0.8 -0.7 -0.4 0.3000 -1.9 -2.1 -1.8 -1.4 -1.4 -0.8 0.4000 -2.7 -2.6 -2.7 -2.3 -2.2 -1.7 0.5000 -3.1 -3.0 -3.1 -3.0 -2.8 -2.3 0.5997 -4.2 -4.1 -4.1 -3.9 -3.5 -3.5 0.7000 -4.7 -4.8 -4.8 -4.4 -4.4 -4.1 0.7994 -4.8 -4.5 -4.4 -4.0 -4.0 -3.8 0.9000 -3.8 -3.6 -3.5 -3.3 -3.1 -3.3 1.0000 0.0 0.0 0.0 0.0 0.0 0.0
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 114 Figure 4.31. Excess molar volume ( V E ) for the binary system [C2mim][OAc] + methanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: : ●, 278.15 K; ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. Figure 4.32. Excess molar volume ( V E ) for the binary system [C2mim][OAc] + ethanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ●, 278.15 K; ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K; □, 328.15 K; , 338.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. x IL 0.0 0.2 0.4 0.6 0.8 1.0 V E / cm 3 ꞏmol -1 -1.8 -1.4 -1.0 -0.6 -0.2 0.2 x 1 0.0 0.2 0.4 0.6 0.8 1.0 V E / cm 3 ꞏmol -1 -1.8 -1.4 -1.0 -0.6 -0.2 0.2
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 115 Figure 4.33. Excess molar volume ( V E ) for the binary system [C2mim][OAc] + 1-propanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K; □, 328.15 K; , 338.15 K; , 348.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. Figure 4.34. Excess molar volume ( V E ) for the binary system [C2mim][OAc] + 2-propanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K; □, 328.15 K; , 338.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. x IL 0.0 0.2 0.4 0.6 0.8 1.0 V E / cm 3 ꞏmol -1 -1.6 -1.2 -0.8 -0.4 0.0 x IL 0.0 0.2 0.4 0.6 0.8 1.0 V E / cm 3 ꞏmol -1 -1.6 -1.2 -0.8 -0.4 0.0
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 116 The series of viscosity logarithm change of mixing at constant temperature are depicted as a function of composition in Figures 4.35 to 4.38 for each binary system. It can be observed that, for any of the systems, the viscosity logarithm change of mixing is positive over the entire range for any of the isotherms experimentally investigated, and it increases with increasing temperature for any given composition. This means that the viscosity of the mixtures is higher than that predicted by the mixing rule by Arrhenius and Kendall (Equation 4.14). All systems exhibit maxima of Δln( / 0) shifting from xIL ≈ 0.5-0.6 at low temperatures to xIL ≈ 0.3-0.4 at high temperatures. For the molar refraction change of mixing, unfortunately, the relatively small values obtained and reported in Tables 4.10 to 4.13 precluded the possibility of getting sufficiently smooth trends as to solidly discuss any composition effect in this magnitude for the studied systems. Figure 4.35. Viscosity logarithm change of mixing (Δln( η / η 0)) for the binary system [C2mim][OAc] + methanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ●, 278.15 K; ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. x IL 0.0 0.2 0.4 0.6 0.8 1.0 ln ( 0.0 0.2 0.4 0.6 0.8
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 117 Figure 4.36. Viscosity logarithm change of mixing (Δln( η / η 0)) for the binary system [C2mim][OAc] + ethanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ●, 278.15 K; ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K; □, 328.15 K; , 338.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. Figure 4.37. Viscosity logarithm change of mixing (Δln( η / η 0)) for the binary system [C2mim][OAc] + 1propanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K; □, 328.15 K; , 338.15 K; , 348.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. x IL 0.0 0.2 0.4 0.6 0.8 1.0 ln ( 0.0 0.2 0.4 0.6 0.8 x IL 0.00.20.40.60.81.0 ln ( 0.0 0.2 0.4 0.6 0.8
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 118 Figure 4.38. Viscosity logarithm change of mixing (Δln( η / η 0)) for the binary system [C2mim][OAc] + 2propanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ○, 288.15 K; ▼, 298.15 K; , 308.15 K; ■, 318.15 K; □, 328.15 K; , 338.15 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. The surface tension change of mixing is plotted, as a function of composition, for the different isotherms investigated, in Figures 4.39 to 4.42 for all four studied binary systems. As it can be observed in Figure 4.39, the binary system [C2mim][OAc] + methanol presents values of ∆ clearly positive throughout the entire composition range, with maxima in the band of composition xIL = 0.3-0.5 for all the isotherms. In the case of the system [C2mim][OAc] + ethanol (Figure 4.40), the smoothness of the series is not as good as it would be desired, likely due to the fact that the uncertainty of the calculated values of ∆ in this case is relatively higher in comparison to their magnitude. No particular smooth trend with composition was found, nor an evident evolution with temperature for a given composition, although general trends can still be intuited. Thus, in general there seems to be a positive maximum for the mixtures rich in ethanol, which would imply that in these mixtures the surface of the liquid tends to get enriched in the ionic liquid (which is the pure compound with the highest surface tension) with respect to the bulk composition. And for the evolution with temperature, a shift towards less negative/more positive values with increasing temperature is loosely apparent, which could be connected with a higher concentration of ionic liquid at the surface of the liquid mixture promoted by a rise in temperature. Regarding the system xIL 0.00.20.40.60.81.0 ln ( 0.0 0.2 0.4 0.6 0.8
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 119 [C2mim][OAc] + 1-propanol (Figure 4.41), again problems to identify smooth trends and evolutions were noticeable, obtaining a plot of series with quite scattered data. Nevertheless, it can still be stated that its values of ∆ are typically negative and the different isotherms reach a minimum (a maximum in absolute value) in the band of mole fraction of ionic liquid xIL = 0.60-0.90 for all the temperatures investigated. In the propanol-rich region, the ∆ values are generally less negative, being even slightly positive for some of the data points at the higher temperatures, resembling somehow the behaviour described above for the system [C2mim][OAc] + ethanol. In contrast, the system [C2mim][OAc] + 2-propanol (Figure 4.42) shows values of ∆ exclusively negative, with its absolute values being greater (i.e., ∆ more negative), concomitantly leading to a less scattered plot since the uncertainty in this case is smaller in relation to the magnitude of the property change of mixing. Nevertheless, even in the case of this system it is hard to elucidate a clear evolution of ∆ with temperature, as all the isotherms are too close to each other. Figure 4.39. Surface tension change of mixing ( Δ ) for the binary system [C2mim][OAc] + methanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ●, 278.2 K; ○, 288.2 K; ▼, 298.2 K; , 308.2 K; ■, 318.2 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. x IL 0.00.20.40.60.81.0 / mNꞏm -1 0.0 1.0 2.0 3.0 4.0 5.0 6.0
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 120 Figure 4.40. Surface tension change of mixing ( Δ ) for the binary system [C2mim][OAc] + ethanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ●, 278.2 K; ○, 288.2 K; ▼, 298.2 K; , 308.2 K; ■, 318.2 K; □, 328.2 K; , 338.2 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. Figure 4.41. Surface tension change of mixing ( Δ ) for the binary system [C2mim][OAc] + 1-propanol as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ○, 288.2 K; ▼, 298.2 K; , 308.2 K; ■, 318.2 K; □, 328.2 K; , 338.2 K; , 348.2 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. x 1 0.0 0.2 0.4 0.6 0.8 1.0 / mNꞏm -1 -0.8 -0.2 0.4 1.0 1.6 2.2 xIL 0.0 0.2 0.4 0.6 0.8 1.0 / mNꞏm-1 -6.0 -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 121 Figure 4.42. Surface tension change of mixing ( Δ ) for the binary system [C2mim][OAc] + 2-propanol, as a function of the mole fraction of [C2mim][OAc] ( x IL ) at different temperatures: ○, 288.2 K; ▼, 298.2 K; , 308.2 K; ■, 318.2 K; □, 328.2 K; , 338.2 K. Solid lines represent the corresponding correlations by Redlich-Kister polynomials. The evolution of the excess molar volumes, viscosity logarithm change of mixing, and surface tension change of mixing with the composition of the mixtures was correlated, at all the investigated temperatures, by means of Redlich-Kister polynomials (Equation 4.26). Polynomials with m = 3 were found to adequately correlate the series of VE and ln( / 0) in the studied systems, whereas polynomials with m = 2 were selected for ., as they already produced sufficiently low values of the root mean square deviations (rmsd). The Redlich-Kister coefficients obtained by least-squares fits are listed in Tables 4.14 to 4.17, where the corresponding rmsd values obtained with Equation 4.27 are also included. The corresponding correlation functions are depicted, along with the experimental data points, in Figures 4.31 to 4.42 as solid lines for each of the isotherms experimentally investigated. The quality of these correlations can thus be visually assessed. xIL 0.0 0.2 0.4 0.6 0.8 1.0 / mNꞏm-1 -6.0 -5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 128 confocal microscopy. Thus, in some microscopic images corresponding to transparent solutions, small heterogeneities constituted by solvated portions of biopolymer were observed (Figure 4.46). In any case, for the results and effect of the work reported herein, the criterion of transparent solution was the one on which the “soluble” condition was decided. In Table 4.18, it can be clearly observed that none of the biopolymer standards was appreciably soluble in pure alcohol (a value of solubility lower than 0.1 g of solute per 100 g of alcohol). Conversely, the solubility of MCC was quite high in pure [C2mim][OAc], as compared to the solubility of cellulose in other solvents, or even in other ionic liquids (Wang et al., 2012). This solubility shows a similar trend for both systems [C2mim][OAc] + methanol and [C2mim][OAc] + ethanol: its value decreases moderately with a moderate increase in the percentage of alcohol in the solvent composition (xIL= 0.60, 0.80). The solubility in the system with ethanol is significantly lower than in the system with methanol for a given molar concentration (for example: for a composition xIL= 0.60, the solubility in the system with methanol is 11 g of MCC per 100 g of solvent, whereas in the system with ethanol it is 4 g of MCC per 100 g of solvent). With a further increase in the concentration of alcohol (xIL= 0.40 or lower), MCC became practically insoluble (less than 0.5 g per 100 g of solvent in any case). For xylan, a lower solubility than that of MCC in the pure ionic liquid was found. Interestingly, this solubility decreases much more rapidly if methanol is added than if ethanol is added. In fact, for a solvent composition with xIL= 0.80, in the case of ethanol the solubility of xylan was still equivalent to that of the pure ionic liquid, whereas for the case of methanol the solubility of xylan was already negligible at that composition (<0.1 g per 100 g of solvent). A slightly higher solubility of xylan (although still very low: in the range 0.5-1.0 g per 100 g solvent) was experimentally found in the [C2mim][OAc] + methanol mixture with xIL= 0.20 (corresponding to wIL= 0.57). This non-monotonic evolution of the solubility with composition was previously reported for the case of biopolymers in mixed solvents; for example in the case of lignin in a mixture with acetone and water (Sun et al., 2011). It is not improbable that there be a local maximum of solubility of xylan in the mixtures of methanol and [C2mim][OAc] in the region relatively rich in methanol; but, due to the low solubilities in absolute terms, this aspect was not further investigated. Regarding the solubility of Indulin AT, it was practically negligible in pure alcohol (<0.1 g per 100 g of solvent), but very large (>20 g per 100 g of solvent) in the solvent combination tested: xIL= 0.20, 0.40, and 0.60;
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 129 whereas for xIL= 0.80 or 1.00, the system became like a gel and its handling turned impracticable in the process of solubilising the large amounts of solute added. Such large solubility looks consistent with the dissolution of up to 300 g of Indulin AT per 1 kg of [C2mim][OAc] reported in the literature at a temperature of 363 K (Lee et al., 2009). In trying to refine the abrupt transition from negligible solubility in pure alcohol to a solubility higher than 20 g of Indulin AT per 100 g of solvent in the mixture of [C2mim][OAc] with alcohol (methanol or ethanol) with xIL= 0.20, solubility tests were carried out in mixtures of ionic liquid and alcohol in that range with a step of 0.01 in mole fraction. It was found that a sharp solubility boost occurred at concentrations of ca. xIL= 0.08 for the system [C2mim][OAc] + methanol (equivalent to an ionic liquid mass fraction wIL= 0.32), and ca. xIL= 0.15 for the system [C2mim][OAc] + ethanol (equivalent to wIL= 0.39). The explanation of this behaviour is not clear at present, but it is remarkable to see that it is observed for both systems. A solubility test was also carried out for MCC in the pure ionic liquid at higher temperature of 358 K, more in connection with other cellulose solubility experiments in the literature. A rise in the experimental solubility up to 30 g of MCC per 100 g of [C2mim][OAc] was observed. This value is higher than others reported in the literature for other sources of cellulose (e.g., eucalyptus prehydrolysis sulphate pulp (Eu-569), cellulose from Trichodermareesei, or different types of Avicel) at the same or similar temperatures (Kosan et al., 2008; Zhao et al., 2008; Balensiefer et al., 2008; Vitz et al., 2009; Zavrel et al., 2009; Fu et al., 2010), perhaps due to a lower degree of polymerisation of the cellulose standard used herein. In a hypothetical case of dissolution of biomass, with all three biopolymers involved, the individual solubility values reported in Table 4.18 are expected to be affected by the presence of the other biopolymers in the solution. Nevertheless, those individual solubilities provide a reasonable estimate of the compositions of the solvent system [C2mim][OAc] + alcohol for which each specific biopolymer will precipitate out of the solution or not. Thus, the composition of the solvent system could be conveniently adjusted, for example at different stages in a process, to dissolve one or several of the biopolymers, according to the interest in each particular case. In view of the values in Table 4.18, the pure ionic liquid will dissolve all three biopolymers (up to a certain limit) and the pure alcohol will not dissolve any of them; and, in between, there exist solvent composition ranges in which the carbohydrates and lignin will be codissolved simultaneously, or in which only lignin will dissolve. In a context of
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 130 fr actionation of biomass, these results suggest the possibility of using the [C2mim][OAc]+ alcohol system to separate, via selective dissolution, a mixture of the major lignocellulose biopolymers. Figure 4.47 shows a possible scheme for the fractionated recovery of the biopolymer fractions from “disengaged biomass” (an idealisation of a lignocellulosic biomass previously treated so that all the lignocellulosic bonds linking the biopolymers to each other were fully broken), exemplified for the solvent system [C2mim][OAc] + methanol. Figure 4.47. Proposed process scheme for the fractionation of a mixture of the major lignocellulose biopolymers (“disengaged biomass”) via selective dissolution with different combinations of [C2mim][OAc] (IL) and methanol. The ratios indicated for the IL + methanol mixtures are in a molar basis. An experiment was attempted to verify the validity of the scheme in Figure 4.47, using a mixture of 0.25 g of each of the biopolymer standards (for a total of 0.75 g of mixture). This mixture was first combined with 5.0 g of a 20:80 mol/mol mixture of [C2mim][OAc] + methanol (which is only capable of dissolving lignin appreciable at that concentration). Upon vigorous stirring for several hours and subsequent filtration (under soft vacuum using a fritted glass Allihn filter tube), the remaining solid was combined with 5.0 g of a 60:40 mol/mol mixture of [C2mim][OAc] + methanol (which exhibits a negligible solubility capacity for xylan, but can dissolve cellulose in appreciable levels). The filtration in the second stage turned out to be difficult due to very high viscosity, and therefore only the first separation stage in Figure 4.47 was effectively carried out in the experiment. By elimination of solvents from the separated fractions, 0.14 g of a black precipitate was obtained from the filtrate of the first stage
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 131 (‘Precipitate-1’), and 0.61 g of a brown precipitate was obtained from the mixture in the second stage (‘Precipitate-2’). Photographs of ‘Precipitate-1’ and ‘Precipitate-2’ are shown in Figure 4.48a. The FT-IR spectra of both samples, as well as of those of the equivalently regenerated biopolymer standards, are shown in Figure 4.48b (zoomed in the wavenumber range 500-2000 cm-1). The similarity between the spectra of ‘Precipitate-1’ and Indulin AT is evident, as it is between the spectra of ‘Precipitate-2’ and of MCC and xylan. Among the set of characteristic vibration bands (Casas et al., 2013; Labbé et al., 2005), specific analysis of signals at 897 cm-1 (characteristic of carbohydrates) and at 1459 cm-1 and 1510 cm-1 (characteristic of lignin) in the obtained spectra reveals in this particular case the preferential presence of Indulin AT in ‘Precipitate-1’ and the preferential presence of carbohydrates (MCC and xylan) in ‘Precipitate-2’. This is in good agreement with what would be expected from the colourations observed for the precipitates (Figure 4.48a). Figure 4.48. a) Photographs of ‘Precipitate-1’ (top) and ‘Precipitate-2’ (bottom). b) FT-IR spectra, in the wavenumber range 500-2000 cm-1, of regenerated biopolymer standards and of regenerated precipitates of the selective dissolution experiment (see details in the text); from top to bottom: MCC, xylan, Indulin AT, ‘Precipitate-1’, and ‘Precipitate-2’. The solid vertical line indicates a characteristic signal of cellulose/hemicellulose, and the dashed vertical lines indicate characteristic signals of lignin (Casas et al., 2013; Labbé et al., 2005).
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 132 4.4.4.Precipitationtests:thealcoholasantisolvent On the basis of the differences in the solubilities reported in Table 4.18 for the biopolymer standards in mixtures of [C2mim][OAc] + alcohol of varied compositions, a precipitation scheme could be considered for the fractionated recovery of biopolymer fractions from lignocellulosic biomass dissolved in [C2mim][OAc] by gradual additions of controlled amounts of alcohol acting as antisolvent. In order to evaluate the ability of the alcohol to act as antisolvent for this fractionated recovery, some precipitation tests were carried out. In particular, the precipitation of the biopolymer standards, previously dissolved in [C2mim][OAc], by addition of methanol was investigated. Solutions of 0.15 g of biopolymer standard (MCC, xylan, or Indulin AT) in 3.00 g of [C2mim][OAc] (a solute-to-solvent mass ratio of 5:100) were made and subsequently subjected to the addition of different amounts of methanol at ambient temperature. Although the values reported in Table 4.18 would suggest the precipitation of at least the MCC and the xylan with the addition of moderate amounts of methanol (less than 1:1 wt/wt ratio to the ionic liquid), a clean precipitation of solids was not observed in the tests. Instead, a kind of emulsion or gel-like phase was formed (exemplified in the photographs of Figure 4.49 for the case of cellulose and xylan). Similar gel phase formations have been previously reported in the literature for similar systems (Dibble et al., 2011). This fact would hinder the utilisation of the alcohol in moderate amounts as a direct antisolvent for the regeneration of lignocellulose fractions previously dissolved in [C2mim][OAc] or for the fractionated precipitation of the different biopolymers from a [C2mim][OAc] solution by gradual addition of methanol (see scheme in Figure 4.50). This emulsion/gel formation was observed even after the addition of as much as five times the mass of methanol with regard to the ionic liquid (methanol-to-[C2mim][OAc] ratio of 5:1 wt/wt). No large amounts of methanol were attempted to force a clean precipitation, since this would lead to an excessively high amount of volatile organic used in the process, as well as to the consumption of too much energy in the elimination of the antisolvent for recycling of the ionic liquid in the cooking step (see Figure 4.50) in a continuous process.
4. Ionicliquid+alcoholsystems.Solubilityofbiopolymers 133 Figure 4.49. Illustrative pictures from the precipitation tests: a) gel formation upon addition of methanol, in a ratio of 1:1 wt/wt relative to ionic liquid, to a solution of cellulose in [C2mim][OAc]; b) emulsion formed upon addition of methanol, in a ratio of 1:1 wt/wt relative to ionic liquid, to a solution of xylan in [C2mim][OAc]. See the text for further details. Figure 4.50. Potential scheme for the dissolution and fractional precipitation of lignocellulosic biomass using [C2mim][OAc] as solvent and methanol as antisolvent. This scheme would not be viable with the use of moderate amounts of methanol, due to the formation of gel-like phases. Solutionofbiomassin[C 2 mim][OAc] Carbohydrate‐rich fraction Solutionofligninin [C 2 mim][OAc]+methanol [C 2 mim][OAc]+ methanol Milled,unpretreated biomass Lignin Additionof[C 2 mim][OAc] Addition#2ofmethanol Methanol [C 2 mim][OAc] Cooking Distillation Filtration Filtration Addition#1ofmethanol
5. PRETREATMENT OF EUCALYPTUS WOOD WITH AN IONIC LIQUID + ETHANOL MIXTURE
5. Pretreatmentofeucalyptuswoodwithanionicliquid+ethanolmixture 137 5. PRETREATMENTOFEUCALYPTUS WOODWITHANIONICLIQUID+ ETHANOLMIXTURE 5.1.Motivation Much of the research efforts made to date on the pretreatment of lignocellulosic biomass with ionic liquids has put its emphasis on the partial or total dissolution of the biomass in the ionic liquid. Due to the non-volatile nature of both the ionic liquid and the biopolymeric constituents of the biomass, the recovery of the dissolved fractions has typically been addressed by the addition of large amounts of a molecular liquid miscible with the ionic liquid and acting as antisolvent of the lignocelluloses. To avoid the large input of energy associated with the recovery of the ionic liquid from its mixture with the antisolvent by vaporisation of the latter, an alternative pretreatment scheme based on non-dissolving conditions may be envisioned. In a non-dissolving pretreatment, no fractions of lignocellulosic material would be actually dissolved in the pretreatment fluid at relevant levels. Instead, the pretreatment fluid would interact with the biopolymers in the solid phase (without carrying out their dissolution), in a way that would facilitate the reaction and transformation of these biopolymers in subsequent stages. An example of this might be a non-dissolving pretreatment in which the crystallinity of the cellulose fraction in the lignocellulosic biomass is reduced, thus facilitating its reactivity in later steps within the biorefinery scheme. Given the non-dissolving character of this approach, the use of antisolvents to precipitate dissolved fractions would not be necessary, and the reclaim of the pretreated biomass would be doable by simple filtration. In proposing a fluid solvent for the non-dissolving pretreatment, an ionic liquid able to interact with the lignocellulosic biomass or its constituent biopolymers under certain conditions could be an interesting choice. Moreover, the consideration of the mixture of such ionic liquid with a molecular cosolvent capable of conveniently tuning this dissolution capacity, as well as other properties of the solvent, could be also of
AppendixC: Publications
APPENDIXC:Publications 243 AppendixC: Publications The work contained in this thesis has produced, to date, the following research articles published in internationally reputed scientific journals: Castro, C. M.; Rodríguez, H.; Arce, A.; Soto, A. 2014. “Mixtures of Ethanol and the Ionic Liquid 1-Ethyl-3-methylimidazolium Acetate for the Fractionated Solubility of Biopolymers of Lignocellulosic Biomass”, Ind.Eng.Chem.Res., 53, 11850–11861. Castro, C. M.; Arce, A.; Rodríguez, H.; Soto, A. 2015. “Influence of Methanol on the Dissolution of Lignocellulose Biopolymers with the Ionic Liquid 1-Ethyl-3methylimidazolium Acetate”, Ind.Eng.Chem.Res., 54, 9605–9614. Castro, C. M.; Arce, A.; Soto, A.; Rodríguez, H. 2016. “Thermophysical Characterization of the Mixtures of the Ionic Liquid 1-Ethyl-3-Methylimidazolium Acetate with 1-Propanol or 2-Propanol”, J.Chem.Eng.Data, 61, 2299–2310. Castro, C. M.; Arce, A.; Soto, A.; Rodríguez, H. 2016. “Liquid-liquid equilibria of mutually immiscible ionic liquids with a common anion of basic character”, J.Chem. Thermodyn., 102, 12–21. Two more manuscripts are in preparation at the moment of submitting this thesis, involving the work presented in Chapters 5 and 6. They will be submitted for publication in internationally renowned journals in the next months. Additionally, the following patent application, in part considering work included in this thesis, has been filed: Holding, A. J.; Castro Valiña, M. C.; Parviainen, A.; Kilpeläinen, I.; Rodríguez Martínez, H.; King, A. W. T. 2018. “Method for pre-treating cellulosic material”, Finnish Patent Application Number 201805501.
AppendixD: “Resumen”(Summary, inSpanish)
APPENDIXD:“Resumen”(Summary,inSpanish) 247 AppendixD: “Resumen”(Summary,inSpanish) La búsqueda de procesos más sostenibles, menos contaminantes y con un impacto ambiental reducido es una tendencia de actualidad en el sector industrial. En este contexto, el reemplazo de materias primas no renovables (en las que aún se fundamenta en gran medida el tejido industrial actual) por recursos de carácter renovable podría ser la base sobre la que desarrollar una nueva plataforma química industrial realmente sostenible para la producción de substancias químicas y materiales. Dentro de las opciones de recursos biorrenovables, la biomasa lignocelulósica se produce naturalmente en cantidades suficientes como para poder erigirse en la materia prima de referencia de esa nueva plataforma, dando cobertura a los grandes volúmenes actuales de producción industrial que se encargan de satisfacer nuestras necesidades. Además, la biomasa lignocelulósica presenta la ventaja añadida de estar geodistribuida de manera más homogénea que los recursos primarios no renovables sobre los que se basa la producción industrial hoy en día. Las paredes celulares de las plantas están compuestas mayormente (otras sustancias suelen representar un pequeño porcentaje) por tres biopolímeros: celulosa, hemicelulosa y lignina, de características y naturaleza química dispares, que ofrecen una idea de la riqueza que las lignocelulosas atesoran para su potencial utilización en un marco de biorrefinería. No obstante, los biopolímeros mencionados se organizan en la pared celular de un modo complejo y de fuerte carácter recalcitrante (como consecuencia de la propia evolución de las plantas dirigida a resistir la acción de agentes degradantes). Por esta razón, para el adecuado aprovechamiento industrial de los biopolímeros presentes en el material lignocelulósico se hace necesario realizar lo que se denomina etapa de pretratamiento. Esta etapa tiene por objeto la modificación de la estructura lignocelulósica y la alteración de su tamaño de poro, por ejemplo a través de la reducción de la cristalinidad que presenta la fracción de celulosa en la planta en su estado natural. Como la explotación industrial de la biomasa lignocelulósica como recurso para la industria química se ha centrado históricamente en la producción de celulosa, no
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 248 resulta extraño que los procesos industriales más desarrollados en la actualidad para el tratamiento de este recurso biorrenovable sean los de producción de pasta de celulosa y papel. En estos procesos la etapa de pretratamiento constituye en realidad la esencia del tratamiento, aplicando métodos agresivos que degradan/eliminan la hemicelulosa y la lignina en condiciones de operación severas, y con un importante impacto ambiental. Más allá del sector de pasta de celulosa y papel, se han desarrollado múltiples métodos de pretratamiento de biomasa lignocelulósica de diversa naturaleza: pretratamiento mecánico, hidrólisis con ácido diluido, hidrólisis alcalina, steam explosion, AFEX (a partir de sus siglas del inglés: ammoniafibreexplosion), métodos organosolv, pretratamiento biológico, etc. A pesar de esta variedad, todos estos tipos de pretratamiento presentan inconvenientes significativos, que se traducen en que el coste de pretratamiento en procesos para el aprovechamiento de lignocelulosas suponga un porcentaje muy relevante del coste total. Así, es clara la necesidad de desarrollar mejores procesos de pretratamiento. El concepto de biorrefinería ha emergido con fuerza en los últimos años, tratando de superar el foco en la producción de celulosa que se ha mencionado anteriormente, de manera que se aproveche también el potencial ofrecido por la hemicelulosa y la lignina en la materia prima. A día de hoy está ampliamente aceptado que el desarrollo de una biorrefinería viable pasa por la consideración de la valorización de los tres biopolímeros. En este contexto, se acentúa la necesidad de desarrollo de métodos de pretratamiento alternativos, que sean capaces de alterar la biomasa lignocelulósica, sin degradación de sus biopolímeros constituyentes, de manera que estos mismos se puedan aprovechar óptimamente en etapas de proceso posteriores. Los líquidos iónicos son un tipo de sustancias constituidas íntegramente por iones y que presentan un punto de fusión (o de transición vítrea) relativamente bajo – inferior a 373 K. Estas sales han captado una importante atención del mundo académico e industrial durante los últimos años, fundamentalmente a causa de un interesante conjunto de propiedades que muchos líquidos iónicos poseen: presión de vapor despreciable, relativamente buenas estabilidades térmica y química, carácter no inflamable, gran capacidad para la disolución de compuestos muy variados, etc. Además, sus propiedades son ajustables a una determinada aplicación hasta un cierto punto mediante la apropiada selección de la combinación catión-anión y el diseño de sus estructuras químicas. Las aplicaciones para las que se han propuesto son
APPENDIXD:“Resumen”(Summary,inSpanish) 249 tremendamente variadas, con algunas aplicaciones ya convertidas en realidad a nivel industrial. Entre las aplicaciones de líquidos iónicos que ofrecen gran potencial cabe destacar su uso en procesos para el tratamiento de materiales (ligno)celulósicos en un marco de biorrefinería. Dada la capacidad de algunos líquidos iónicos para disolver celulosa e incluso lignocelulosas como madera, estos líquidos iónicos pueden constituir la base de nueva tecnología para, por ejemplo, el pretratamiento efectivo de la biomasa lignocelulósica para una mejor valorización de sus biopolímeros constituyentes. Objetivos El objetivo general de esta tesis doctoral es el avance en el conocimiento sobre las posibilidades y el potencial de sistemas de fluidos basados en líquidos iónicos para la configuración de mejores esquemas de proceso para el pretratamiento de biomasa lignocelulósica. Este objetivo general incluye objetivos específicos con diferentes sistemas fluidos y esquemas de proceso, desde un enfoque en la ciencia fundamental que rige los sistemas investigados hasta su utilización en enfoques más aplicados. Así, un primer sistema fluido de interés es el constituido por dos líquidos iónicos mutuamente inmiscibles, uno con capacidad para disolver lignocelulosas y el otro no, y cuya miscibilidad mutua sea variable con un cambio en la temperatura. También es un objetivo la investigación del potencial de alcoholes ligeros como codisolventes o antidisolventes del más paradigmático líquido iónico para disolver biomasa hasta la fecha: acetato de 1-etil-3-metilimidazolio; así como la caracterización térmica y física de estos sistemas binarios. En un plano más aplicado, se busca que un sistema fluido de este último tipo pretrate satisfactoriamente madera (Eucalyptusglobulus) en condiciones de baja temperatura (sin que se produzca disolución, con especial atención a la posible reducción en el grado de cristalinidad de la madera). En una línea similar, esta tesis también pretende explorar el uso de líquidos iónicos con otro tipo de catión (concretamente tetraalquilfosfonio) para el pretratamiento directo de madera (Picea abies, i.e. abeto noruego), con o sin codisolventes, así como diferentes variables de proceso y su influencia en el grado de fibrilación de las partículas y la composición de la madera pretratada resultante.
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 256 Pretratamientodemaderadepíceaconlíquidosiónicosde fosfonio Se estudió el pretratamiento de fragmentos de astilla de Piceaabies, así como de serrín industrial de la misma especie, mediante contacto con dos líquidos iónicos de acetato de tetraalquilfosfonio con cadenas sustituyentes de alquilo de diferentes longitudes: acetato de tetrabutilfosfonio ([P4 4 4 4][OAc]) y acetato de metiltrioctilfosfonio ([P8 8 8 1][OAc]). Ninguno de estos líquidos iónicos presenta capacidad para la disolución de madera en estado puro, si bien se sabe que pueden disolver celulosa en combinación con algún disolvente polar aprótico como el dimetilsulfóxido. En un conjunto inicial de experimentos, se exploraron como disolventes de pretratamiento [P4 4 4 4][OAc], dimetilsulfóxido y -valerolactona (otro disolvente polar aprótico, y además de origen biorrenovable) en estado puro, así como las mezclas a un 50 % en peso del líquido iónico con cada uno de los disolventes moleculares. Se combinaron estos disolventes con las astillas en una relación de 5 g de biomasa por cada 100 g de disolvente, y se efectuaron los pretratamientos por un período de ca. 16 h a diferentes temperaturas en el rango 353-413 K. Los resultados arrojaron poco efecto o degradación térmica evidente en los casos en los que el dimetilsulfóxido o la -valerolactona participaban en el disolvente pretratante. En el caso del [P4 4 4 4][OAc] puro, a las temperaturas más elevadas se observó claramente fibrilación en la madera pretratada. En vista del efecto negativo de la presencia de los disolventes moleculares en los experimentos precedentes, en el caso del [P8 8 8 1][OAc] se realizaron experimentos únicamente con el líquido iónico puro. También se observó fibrilación, aunque más incipiente y prácticamente sólo a la temperatura de 413 K, la más alta de las exploradas. Analizando la fase líquida tras el tratamiento, se pudieron identificar pequeñas trazas asociables a los tres biopolímeros principales de la biomasa, quizás en mayor medida de fracciones hemicelulósicas; pero en cualquier caso cantidades muy pequeñas, confirmando el carácter de no-disolución del pretratamiento propuesto. El análisis composicional de la madera pretratada indicó una disminución relativa de la composición de hemicelulosa, siendo justamente esta disminución más relevante en el caso de pretratamiento con [P4 4 4 4][OAc], que es el que condujo a un pretratamiento más eficaz.
APPENDIXD:“Resumen”(Summary,inSpanish) 257 Conclusiones A continuación se enumeran una serie de conclusiones que se derivan de los resultados originales presentados en esta tesis para diferentes aproximaciones de proceso utilizando sistemas fluidos basados en líquidos iónicos para el pretratamiento de biomasa lignocelulósica: Los equilibrios líquido-líquido de líquidos iónicos mutuamente inmiscibles desarrollados en esta tesis vienen a ampliar el corpus de conocimiento de este tipo de sistemas, muy limitado en la bibliografía científica. En el contexto de su potencial utilización como sistemas fluidos integrados disolvente-antidisolvente en el pretratamiento de materiales lignocelulósicos, desafortunadamente no se alcanzan puntos UCST o LCST dentro del rango de temperaturas en el que estas mezclas se comportan como líquidos estables. No obstante, dado el carácter novolátil de los líquidos iónicos y su buena estabilidad térmica, estos sistemas bifásicos líquido-líquido pueden ser de particular interés en extracciones a alta temperatura. La densidad, viscosidad, índice de refracción y tensión superficial de las mezclas binarias [C2mim][OAc] + (metanol, etanol, 1-propanol o 2-propanol) disminuyen con un aumento de la temperatura o un aumento de la concentración de alcohol. La viscosidad es correlacionable con la temperatura, para una composición dada, mediante la ecuación de Vogel-Fulcher-Tamman (excepto en el caso de los alcoholes puros, para los que funciona mejor la ecuación de Andrade); mientras que el resto de propiedades pueden ser correlacionadas adecuadamente mediante líneas rectas o, a lo sumo, un polinomio de segundo grado (para el caso de la densidad). El volumen molar de exceso en estas muestras evidencia la predominancia de fuerzas atractivas entre el líquido iónico y el alcohol. No obstante, el [C2mim][OAc] es recuperable de sus mezclas con alcohol mediante la vaporización de este último a temperaturas inferiores a la de degradación del líquido iónico. La capacidad de disolución de biopolímeros por parte de estos sistemas fluidos se puede regular mediante la concentración. No obstante, la precipitación fraccionada de biopolímeros disueltos en [C2mim][OAc] mediante adición de alcohol puede presentar complicaciones debido a la formación de emulsiones y geles.
Fluidsystemsbasedonionicliquidsforthepretreatmentoflignocellulosicbiomass 258 El pretratamiento de partículas de Eucaliptusglobulus con [C2mim][OAc] o con su mezcla con etanol (5 % en peso de este último) a una temperatura de 338 K conduce a un alto grado de fibrilación y una importante reducción de la cristalinidad de la madera en condiciones de no-disolución. Este resultado debería facilitar el procesamiento y transformación de los biopolímeros de la madera en siguientes etapas de proceso. El pretratamiento de astillas de Piceaabies con [P4 4 4 4][OAc] a 413 K resulta en una fibrilación sustancial de la madera. No obstante, la utilización de dimetilsulfóxido o -valerolactona como co-disolventes del líquido iónico conduce a malos resultados de pretratamiento. Algo de fibrilación también se consigue en el pretratamiento con [P8 8 8 1][OAc], pero en menor medida que con [P4 4 4 4][OAc]. La madera pretratada presenta una concentración relativamente menor de hemicelulosa en comparación a la madera original.